Journal of Photochemistry and Photobiology B: Biology 130 (2014) 226–233

Contents lists available at ScienceDirect

Journal of Photochemistry and Photobiology B: Biology journal homepage: www.elsevier.com/locate/jphotobiol

Synergistic antimicrobial activity based on the combined use of a gemini-quaternary ammonium compound and ultraviolet-A light Akihiro Shirai a,⇑, Mutsumi Aihara b, Akira Takahashi b, Hideaki Maseda a, Takeshi Omasa a a Department of Biological Science and Technology, Biosystems Engineering, Institute of Technology and Science, The University of Tokushima, 2-1 Minamijosanjima-cho, Tokushima 770-8506, Japan b Department of Preventive Environment and Nutrition, Institute of Health Biosciences, The University of Tokushima, 3-18-15 Kuramoto-cho, Tokushima 770-8503, Japan

a r t i c l e

i n f o

Article history: Received 1 October 2013 Received in revised form 22 November 2013 Accepted 27 November 2013 Available online 4 December 2013 Keywords: Disinfection Bactericidal action Gemini-quaternary ammonium salt Ultraviolet-A Reactive oxygen species

a b s t r a c t This study examined the utility of synergistic disinfection employing a gemini-quaternary ammonium salt (a gemini-QUAT, namely 3,30 -(2,7-dioxaoctane)bis(1-decylpyridinium bromide)), as an organic biocide in combination with irradiation by an ultraviolet-A (UV-A) light-emitting diode (LED) with a peak wavelength of 365 nm. The combined system represents a novel disinfection method utilizing facilitated in situ oxidation depending on overproduction of reactive oxygen species (ROSs) triggered by the initial action of the gemini-QUAT on the bacterial membrane. We demonstrate that this combination decreased the viability of pathogenic bacteria in a significant and rapid manner, and depended on doses of the gemini-QUAT and the fluence: the viability of Escherichia coli was reduced by greater than 5.0-logs by the combination procedure, but the decrease in viability was only 2.3-logs for exposure to UV at the same fluence dose in the absence of the gemini-QUAT. Adding catalase as a radical scavenger decreased bacterial inactivation by the combined disinfection procedure. Flow cytometric analysis indicated superoxide and hydrogen peroxide overproduction within cells treated with the combined disinfection procedure. The excessive superoxide, detected only in the combined system, appeared to be generated by the action of the gemini-QUAT at the bacterial membrane, leading to excessive and rapid generation of ROS in the system. Our data strongly suggested that this ROS promoted bacterial membrane peroxidation during initial treatment by the combination method, resulting in increased oxidative modification of DNA. These oxidative reactions may play an important role in the efficacy of this disinfection procedure. Ó 2013 Elsevier B.V. All rights reserved.

1. Introduction Irradiation with ultraviolet-C (UV-C) light traditionally has been used as an effective disinfection procedure. Applications of UV-C irradiation have been employed for disinfection of water, air, and surfaces [1]. Generally, the germicidal potency of UV occurs at a primary wavelength of 254 nm. UV-C at 254 nm facilitates the formation of cyclobutane pyrimidine dimers from adjacent thymidines in DNA. On the other hand, it has been reported that irradiation with ultraviolet-A (UV-A) light at wavelengths of 315–400 nm (particularly light with wavelength of 365 nm) has lethal effects on bacteria, including pathogenic microbes [2–5]. The high bactericidal efficacy of 365-nm UV-A irradiation results from the generation of reactive oxygen species (ROS) such as hydrogen peroxide and hydroxyl radical [3]. It is believed that the ROS generation initiates critical bactericidal events, including the oxidation of lipid membranes [6], DNA [3] and intracellular proteins [7]. However, disinfection by UV-A light alone requires ⇑ Corresponding author. Tel.: +81 886567519. E-mail address: [email protected] (A. Shirai). 1011-1344/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jphotobiol.2013.11.027

light sources that emit 365-nm light at high fluence rate to inactivate bacteria effectively. With lower fluence rate sources, bacteria exhibit tolerance to UV-A irradiation. For example, Pseudomonas aeruginosa induced a relA-dependent adaptive response following low fluence rate (2.8 mW/cm2 at 365 nm) UV-A irradiation, rendering this bacterium resistant to further irradiation [8]. Thus, in order to inactivate bacteria rapidly and consistently regardless of the fluence rate of the UV-A light source, some combined disinfectant systems have been constructed. Similar combined applications are used with antibiotics; for instance, co-administration of two antibiotics provides synergistic antimicrobial activity in vitro against clinical isolates of fungi [9]. In an analogous fashion, the bactericidal activity of UV-A light is enhanced in combination with titanium dioxide particles [10,11], with particles of titanium dioxide and silver [12], or with hydrogen peroxide [13]. Increase of the bactericidal activity in the combined systems was attributed to the enhanced oxidative reactions caused by extra-cellular ROS generated by the photocatalysis/photolysis generated by the combined materials. The combination of titanium dioxide and low fluence rate (0.8 mW/cm2 at 365 nm) UV-A light provided synergistic elevation of bactericidal activity compared to either

A. Shirai et al. / Journal of Photochemistry and Photobiology B: Biology 130 (2014) 226–233

treatment alone, leading to a 96% loss of viability following 30 min of treatment [11]. However, in order to obtain higher bactericidal potency and to permit larger-volume treatment, time-consuming and difficult post-treatment procedures (e.g., separation of insoluble titanium dioxide after the disinfectant treatment) are required. High bactericidal activities were achieved by disinfectant systems combining UV with soluble materials such as hydrogen peroxide, but bacterial suspensions containing hydrogen peroxide had to be irradiated with high fluence rate (110 mW/cm2 at 400 ± 20 nm) light to provide decomposition of hydrogen peroxide to higher oxidation states by photoreaction [13]. The disadvantages apparent in these examples imply the need to develop more effective and convenient combined disinfection procedures. Advances in these methods would facilitate easier sterile manufacturing; reductions in the required total fluence doses would counteract limitations such as the high cost of light source devices and the low potency of light-permeation resulting from the presence of particulates and contaminants in treated water. For example, 0.1% riboflavin (a photosensitive water-soluble organic substance that is ‘‘generally regarded as safe’’) has been used in combination with UV-A irradiation with a low fluence rate (3 mW/cm2 at 365 nm) to treat Acanthamoeba [14], suggesting a promising photochemical therapy for Acanthamoeba keratitis. In the present study, we constructed a new disinfection system that combines a soluble organic biocide (a gemini-quaternary ammonium compound (gemini-QUAT; 3,30 -(2,7-dioxaoctane)bis(1-decylpyridinium bromide)), also referred to as 3DOBP4,10) with UV-A irradiation. Gemini-QUATs are composed of two similar QUAT moieties linked by hydrocarbon chains with or without heteroatoms. Such gemini-QUATs have greater surface activities [15], wider antimicrobial spectra, and more potent antimicrobial activities than conventional QUATs, such as benzalkonium chloride [16–18]. Moreover, 3DOBP-4,10 has two advantages: 3DOBP-4,10 has lower toxicity to human cells than conventional QUATs such as benzalkonium chloride, and the compound has high water-solubility (589 g/100 g H2O) [17]. We hypothesized that the powerful antimicrobial activity of 3DOBP-4,10 permits increased entry of oxygen into cells following the interaction of the gemini-QUAT with the cell wall and membrane, resulting in the generation of ROS in the cells [19]. 3DOBP-4,10 appears to synergize with the bactericidal activity of the ROS generated by UV-A light. Unlike some combined systems described previously, the combination of gemini-QUAT + UV-A is expected to generate ROS as an intracellular (not extracellular) species. Fluorescence probes that respond to intracellular ROS demonstrated that an ordinary QUAT, cetyltrimethylammonium bromide (CTAB), produced ROS in cells treated with inhibitory concentrations of CTAB [20]. The purpose of the present study was to demonstrate the synergistic bactericidal activity of the combination of low concentration 3DOBP-4,10 and low fluence rate UV-A irradiation. The high bactericidal activity associated with the combined treatment was also investigated mechanistically by focusing on the associated generation and action of ROS.

227

Fig. 1. Chemical structure of 3DOBP-4,10, a gemini-QUAT used in the combination disinfection process described here.

constructed with 27 LED elements (NCSU033A; Nichia Corp., Anan, Japan) arranged in a three-column-by-nine-row array on a basal plate of 100 mm  300 mm. The fluence rate of the light (wavelength range: 350–385 nm; peak wavelength: 365 nm) emitted by this device was 4.77 mW/cm2 (Fig. 2). The spectrum and light fluence rate were measured at 65 mm from the device, using a cumulative UV meter (MCPD-3700A; Otsuka Electronics Co., Ltd., Hirakata, Japan) that was tested while shielded by the cover of a plastic Petri dish, consistent with the LED’s use in all bactericidal tests. The fluence rate was expressed in mW/cm2 and the total fluence dose, which was calculated based on the fluence rate and exposure time, was expressed in J/cm2. The distance (65 mm) was applied as the irradiation distance between the UV-A–LED device and the surface of bacterial suspensions through all bactericidal assays. In the assays, UV-A irradiation was performed in a dark room, using bacterial suspensions in plastic Petri dishes or six-well culture plates that were held at 30 or 37 °C in a water bath. 2.3. In vitro assay for determination of bactericidal level Bactericidal level was determined by counting of the number of colony-forming units (CFUs). The bacteria used were Escherichia coli NITE Biological Resource Center (NBRC) 12713, Salmonella enterica NBRC 13245, Serratia marcescens NBRC 12648, Staphylococcus aureus NBRC 12732, S. epidermidis NBRC 12993, P. aeruginosa American Type Culture Collection (ATCC) 10145, and S. aureus ATCC 700699 (MRSA). All bacteria were purchased from ATCC and NBRC. After cultivation (15 h, 37 °C) in Luria-Bertani medium (LB broth, Lennox; Nacalai Tesque Inc., Kyoto, Japan), bacterial cells were harvested by centrifugation (6570g, 3 min, 4 °C) and washed twice with the phosphate-buffered saline (D-PBS(–); Nacalai Tesque Inc.). Bacterial suspensions (2.0  106 cells/ml) were prepared in sterile ion-exchanged water, and transferred to a plastic Petri dish (90 mm  20 mm; Nissui Pharmaceutical Co., Ltd., Tokyo, Japan), at 20 ml/plate. A stock solution of 3DOBP-4,10 formulated in sterile milliQ water was added to the bacterial suspension by diluting by a minimum of 1000-fold or more to the prescribed concentration before UV-A irradiation. Following irradiation, aliquots (0.2 ml) of each suspension were subjected to 10-fold serial dilutions with 0.8% (w/v) physiological saline containing 0.7% (w/w) Tween 80 (Kanto Chemical Co., Inc., Tokyo, Japan). In this context, the Tween 80 serves as an inactivating agent that quenches the

2. Materials and methods 2.1. Reagents A gemini-QUAT, 3DOBP-4,10, was synthesized in our laboratory [17]. 3DOBP-4,10 shows no absorption of UV-A light in the wavelength range used in the present study (Fig. 1). 2.2. Light source and irradiation A light-emitting diode (LED) device manufactured by SAN Electronics Ltd. (Aizumi, Japan) was used as a light source. The device is

Fig. 2. Emission spectrum of UV-A–LED used here. This UV-A–LED provides a spectral maximum at 365 nm. The fluence rate reaches 4.77 mW/cm2 at a 65-mm distance between the illumination source and UV meter.

228

A. Shirai et al. / Journal of Photochemistry and Photobiology B: Biology 130 (2014) 226–233

gemini-QUAT, because 10 lM of the QUAT had no effect on the viability of E. coli during 1 h treatment by adding the Tween 80 to its cell suspension (data not shown). Each serial dilution was plated on SCDLP agar (Nihon Pharmaceutical Co., Ltd., Tokyo, Japan) plates and incubated at 37 °C for 36–48 h. Colony counts were determined and used to calculate bactericidal activity, expressed as the log survival ratio, such that the log survival ratio = log (Nt/N0), where N0 is the initial colony count before bactericidal treatment and Nt is the colony count after the treatment for time t. To investigate the effect of ROS on bacterial inactivation using the combination of 3DOBP-4,10 and UV-A irradiation, solutions of superoxide dismutase (from bovine erythrocytes, Cu/Zn type, SOD; Wako Pure Chemical Industries Ltd., Osaka, Japan) and catalase (from bovine liver; Wako Pure Chemical Industries Ltd.) were prepared with D-PBS(), and D()-mannitol (Wako Pure Chemical Industries Ltd.) was dissolved in ion-exchanged water. These antioxidants were added to the bacterial suspensions to final concentrations of 300 U/dish, 20,000 U/dish, and 30 mM, respectively. (Preliminary work (not shown) demonstrated that incubation (1 h, 37 °C, in the dark) with the respective antioxidant resulted in a loss of viability of less than 1.0-log.) The antioxidants at the respective concentrations were pre-incubated with bacteria at 37 °C for 30 min before the UV-A irradiation.

2.4. Analysis of intracellular ROS production by flow cytometry In situ ROS production levels were investigated based on flow cytometry according to a method previously used for mouse fibroblasts [21]. BES-H2O2-Ac (measurement wavelength; kex = 488 nm, kem = 527 nm, Wako Pure Chemical Industries Ltd.) and dihydroethidium (measurement wavelength; kex = 488 nm, kem = 586 nm, Life Technologies Corporation, Carlsbad, CA, USA) were used as specific fluorescence probes responding to ROS produced by the UV-A irradiation with or without 0.1 lM 3DOBP-4,10. The specific fluorescence-labeled E. coli was prepared by gentle shaking (78 rpm) of the bacteria (2.0  107 cells/ml) in D-PBS() containing 30 lM specific fluorescence probe for 80 min at 37 °C in the dark. Cells then were collected by centrifugation (6570g for 3 min at 4 °C) and washed three times with D-PBS(). Bacterial suspensions were adjusted to 2.0  107 cells/ml, distributed at 990 ll per well in six-well culture plates (Zellkultur und Labortechnologie, Switzerland), supplemented with 10 ll of D-PBS() solution with or without 10 lM 3DOBP-4,10, and irradiated with UV-A for 30 or 60 min at 30 °C. The resulting suspensions were recovered, harvested by centrifugation (8000g, 5 min, 4 °C), fixed with 3% (w/v) paraformaldehyde (Sigma–Aldrich Co., LLC., St. Louis, MO, USA) in D-PBS() for 15 min at 37 °C, washed once with D-PBS(), and analyzed in a flow

cytometer (FACS Verse, Becton Dickinson, Franklin Lakes, NJ, USA). As a positive control for each oxidative stress, the bacterial suspension was instead treated (30 min, 30 °C, in the dark) with 0.05% (v/v) hydrogen peroxide or 400 lM tert-butyl hydroperoxide (tBHP), compounds known to induce increased levels of hydrogen peroxide or superoxide [22], respectively. The level of ROS induced by oxidative processes was expressed as arbitrary units (a.u.) normalized to the mean fluorescence values of control samples that were incubated for 30 or 60 min in the absence of both UV-A light and 3DOBP-4,10. 2.5. Measurement of lipid peroxidation induced by bactericidal treatment Lipid peroxidation in E. coli cells was determined by measuring the formation of thiobarbituric acid-reactive substances (TBARS), including aldehydes such as malondialdehyde (MDA). The assay was performed using the Fluorescence TBARS Microplate Assay Kit (Oxford Biomedical Research, Rochester Hills, MI, USA) according to the manufacturer’s instructions [23]. Cell suspensions (2.0  109 cells/ml in D-PBS()) were distributed (995 ll/well) to the individual wells of six-well plates; supplemented to 0 or 50 lM 3DOBP-4,10 by addition of 5 ll D-PBS() solution with or without 3DOBP-4,10; and irradiated with UV-A light for 10 or 60 min at 30 °C. Subsequently, samples of each suspension were spread on SCDLP agar plates to determine viable cell counts. The residual pooled contents of each six-well plate were centrifuged (8000g, 5 min, 4 °C), and the pellets were re-suspended in D-PBS() containing 0.05% (w/v) 2,6-di-tert-butyl-p-cresol (Tokyo Chemical Industry Co., Ltd., Tokyo, Japan) as an antioxidant. Each suspension was sonicated (20 s  4 times at 8 °C) in an ultrasonic bath sonicator (UCT-0312; 27 kHz, 100 W, Tokyo Ultrasonic Engineering Co., Ltd., Tokyo, Japan), and then chromophore fluorescence was measured according to the total MDA procedure of the TBARS assay kit. These fluorescence values were converted to MDA concentrations by comparison to an MDA standard curve, and the extent of lipid peroxidation was expressed as MDA equivalents per 109 cells. 2.6. Measurement of oxidized DNA induced by bactericidal treatment Oxidation of DNA was quantified using a competitive enzymelinked immunosorbent assay (ELISA) kit (High Sensitivity 8-OHdG Check; Japan Institute for the Control of Aging, Nikken SEIL Corp., Fukuroi, Japan). This ELISA kit was based on the detection of 8hydroxydeoxyguanosine (8-OHdG) formed by the oxidative reaction of DNA with ROS [3]. Briefly, bacterial cells treated under the same bactericidal condition as in the TBARS assay were re-suspended in D-PBS() after centrifugation (8000g, 5 min, 4 °C), and each suspension then was processed to extract DNA from bacterial cells (High Pure PCR Template Preparation Kit, Roche Diagnostics K.K., Tokyo, Japan) as described in the manufacturer’s instructions. Subsequently, the decomposition-treatment of 10 lg/150 ll DNA solution was performed according to a method described in the procedure of 8-OHdG Assay Preparation Reagent Set (Wako Pure Chemical Industries Ltd.). Finally, the concentration of 8-OHdG was determined using the ELISA kit. 2.7. Statistical analysis

Fig. 3. Time-course changes in inactivation effect of E. coli after each treatment. The initial cells were suspended at densities of 106 CFU/ml. Symbols: no treatment (in the dark, unshaded square); 0.1 lM 3DOBP-4,10 (unshaded circles); UV-A irradiation alone (shaded triangles); combination of 0.1 lM 3DOBP-4,10 and UV-A irradiation (shaded circles). The data are presented as mean ± SD (n = 3).

All experiments were performed as three to six independent experiments, and the results are presented as the mean with error bars showing the standard deviations. Statistical values (P < 0.05 or 5.0-logs in the absence and presence (respectively) of 0.1 lM 3DOBP-4,10; no loss of viability was seen for exposure to 0.1 lM 3DOBP-4,10 in the absence of UV-A irradiation. These results demonstrated synergism between UV-A and 3DOBP-4,10 in killing this model bacterium. Given the demonstration of activity in E. coli, the efficacy of the combination of UV-A and gemini-QUAT was tested under other conditions and for bacterial pathogens. Again, initial cell densities of 106 CFU/ml were used; lethal concentrations of 3DOBP-4,10 were determined with or without 60 min exposure to UV-A at 17.2 J/cm2 (Fig. 4). The investigation was used to determine the effective minimum concentrations of 3DOBP-4,10 in the combination treatment. If rapid bactericidal ability is desired or lower fluence rate is employed in the combination, water contaminated with bacteria should be treated with a higher concentration of 3DOBP-4,10 than the determined minimum. For E. coli, P. aeruginosa, S. enterica, S. marcescens, S. aureus, and MRSA, survival fell below the detection limit of 10 CFU/ml (that is, >5.0-log reduction in viable cell count) at concentrations of 3DOBP-4,10 ranging from

229

0.05 to 15 lM in the presence of UV-A irradiation. The 0.05) compared to no scavenger sample. 3.3. In situ ROS overproduction by bactericidal process Attenuation of cell killing by catalase treatment confirmed that the bactericidal action of the gemini-QUAT under UV-A light resulted from oxidative stress. Therefore, intracellular ROS levels were estimated by flow cytometric analysis using E. coli cells treated with the combination of 0.1 lM 3DOBP-3,10 and 8.59 J/cm2 of UV-A (30 min irradiation), or with the irradiation alone at 17.2 J/ cm2 of UV-A (60 min irradiation). For these two treatments, bactericidal activity between the treatments did not differ significantly (P = 0.097, based on the results of Fig. 3).

Fig. 4. Dose-dependent effect of 3DOBP-4,10 on inactivation of six strains incubated in the presence (shaded) or absence (unshaded) of total fluence dose at 17.2 J/cm2 of UV-A: A, E. coli; B, P. aeruginosa; C, S. enterica; D, S. marcescens; E, S. aureus; F, MRSA. The initial cells were provided at 106 CFU/ml. Where surviving cell densities were 0.05, Fig. 6C). This experiment demonstrated that a relationship of the superoxide overproduction and 3DOBP-4,10 was very close, because, as described in the previous paper [17], the addition of organic materials to bacteria attenuates the bactericidal action of 3DOBP-4,10. 3.4. Bactericidal treatment-induced lipid peroxidation The TBARS assay was used to investigate excess bacterial lipid membrane peroxidation by the action of ROS. Fig. 7 shows MDA

3.5. Bactericidal treatment-induced DNA oxidation To investigate potential DNA damage, 8-OHdG formation (a marker of DNA oxidation) was measured under the same conditions as used for the TBARS assay. Again, killing efficiencies did not differ significantly between these two treatments (P = 0.14). As shown in Fig. 8, 8-OHdG was detected at 0.51 ng/ml with the combined disinfection treatment system (10 min at 2.87 J/cm2), and at 0.57 ng/ml with UV-A irradiation alone (60 min at 17.2 J/ cm2). The 8-OHdG levels seen with the two treatments did not differ significantly from that seen in a control sample incubated without gemini-QUAT in the dark for 60 min (0.49 ng/ml; P = 0.84 and 0.22 compared to the respective treatment regimens). When the combined disinfection was performed for 60 min at 17.2 J/cm2, 8OHdG was detected at 0.67 ng/ml, a significant elevation compared to the concentrations detected for the combination treatment using 10 min UV-A at 2.87 J/cm2 (P < 0.01) and by 60 min treatment at 17.2 J/cm2 with UV-A alone (P < 0.05). These results demonstrated that DNA oxidation was UV-A-dose dependent in the combination procedure.

4. Discussion The results of this study demonstrated that combining geminiQUAT (an organic biocide) with UV-A irradiation produces synergistic bactericidal activity by increasing intracellular ROS, resulting in elevated oxidation of lipid membranes and DNA. The synergistic activity observed against several bacteria occurred at low fluence rate, in contrast to UV-activated disinfectant systems employing hydrogen peroxide [13], titanium dioxide [11], or riboflavin [24]. A LED was used as the UV-A light source throughout our experiments, instead of a traditional UV mercury lamp or an expensive laser diode. The advantages of LEDs are that these light sources

A. Shirai et al. / Journal of Photochemistry and Photobiology B: Biology 130 (2014) 226–233

Fig. 8. Changes in inactivation level (shaded) and DNA oxidation (unshaded) of E. coli after each treatment. The formation of 8-OHdG by DNA oxidation was quantified using a competitive enzyme-linked immunosorbent assay.  P < 0.05, +P < 0.05, and  P < 0.01 compared to no treatment (in the dark), treatment at 17.2 J/cm2 of UV-A alone, and the combination treatment at 2.87 J/cm2 of UV-A, respectively. ++ P < 0.01 compared to the combination treatment at 2.87 J/cm2 of UV-A, and to treatment at 17.2 J/cm2 of UV-A alone. The data are presented as mean ± SD (n = 3–5).

are not composed of mercury, are not bulky, are durable, and are inexpensive to run. Inactivation efficiency against E. coli was determined at a constant concentration (0.1 lM) of 3DOBP-4,10, and was shown to depend on the total fluence dose of UV-A. A UV-A–LED device previously was reported to demonstrate high bactericidal activity (4.0-log reduction) following 60 min irradiation; the device emitted light (365 nm) of irradiance (70 mW/cm2) sufficient to provide bacterial inactivation without added biocide [3]. In the case of our LED device, no rapid bactericidal inactivation was achieved by 60 min UV-A exposure alone, consistent with the lower fluence rate (4.77 mW/cm2) of our device. The combination of 3DOBP4,10 and exposure to our LED source enhanced bacterial killing compared to the minimal inhibition observed with either treatment alone, suggesting synergistic efficacy of 3DOBP-4,10 in combination with this level of UV-A irradiation. This synergism was confirmed in E. coli and additionally demonstrated for pathogenic bacteria including P. aeruginosa, S. enterica, S. marcescens, S. aureus, and MRSA. This result suggested that the bactericidal activity accelerated by the combination method is based on a synergistic effect by the addition of the gemini-QUAT. The synergistic effect may be useful for the inactivation of several pathogenic bacteria. For example, in a previous report, UV-A light alone at a low fluence rate (3.5 mW/cm2) required 4 h irradiation to reduce S. typhimurium cell viability [2]. In the present work, the addition of 3DOBP-4,10 remarkably strengthened the bactericidal activity by the UV-A irradiation alone during 60 min treatment against five pathogenic bacteria and E. coli. The results suggested that the combined method may find application in sterilizing pathogenic strains under conditions where treatment with either component (gemini-QUAT or UV-A irradiation) alone is insufficient to provide bacterial killing. Moreover, such combinations may be helpful in preventing the emergence of resistance while precluding residual toxicity from the gemini-QUAT, because high activity was obtained by adding an extremely small amount of the QUAT to the UV-A irradiation. Based on the nature of our components, we inferred that the presence of ROS contributed to the synergistic efficacy of our combined treatment procedures. Indeed, our biocide, the gemini-QUAT 3DOBP-4,10, was previously suggested to induce strong antimicrobial activity by inducing intracellular oxidative reactions mediated by ROS [19]. Specifically, the antifungal activity of 3DOBP-4,10 was shown to be attenuated by the presence of SOD (a scavenger of superoxides), of catalase (a scavenger of hydrogen peroxides), or of potassium iodide (a quencher of hydroxyl radicals) [19]. Similarly, previous work demonstrated that both the level of killing and the formation of 8-OHdG (a marker for the formation of an

231

oxidative DNA product) induced by exposure to UV-A were attenuated by the addition of catalase or mannitol (another quencher of hydroxyl radicals) to the treated suspensions [3]. Indeed, in our combined disinfection procedures, the addition of catalase yielded significant attenuation of bactericidal killing efficiency, and that attenuation depended on the use of fresh (non-heat-inactivated) catalase enzyme. Our results implied that hydrogen peroxide was a primary mediator for the synergistic bactericidal activity seen with our combination procedure. The effect of SOD, which quenches superoxide, on survival was not significant. This may be due to the lower oxidizability of superoxide, which fails in effective inactivation, in comparison with other ROS. Role for hydroxyl radicals was not supported by attempts to selectively deplete hydroxyl radicals by D()mannitol or glutathione. Although the inhibition of oxidative reactions by hydroxyl radical, which is the strongest oxidant among ROS, was expected after the addition of D()-mannitol, it did not affect survival in the combination at all. The same phenomenon also occurred with the addition of 1.0 mM glutathione of reduced form in the combination (data not shown). The hydroxyl radical is known to act at very short distances (within approximately 20 nm) owing to this ROS’ short half-life period (approximately 70 ns), inducing lethal oxidative damages on lipid membranes, proteins, and DNA. In our experiment, the rapid breakdown of the radical may preclude detection of attenuation following scavenging with D()-mannitol and glutathione. Nonetheless, we hypothesize the efficacy of our combined disinfection system involves not only hydrogen peroxide but also hydroxyl radical, with additional hydroxyl radical generated from hydrogen peroxide by the Harber–Weiss reaction, including the Fenton reaction. Oxidative stress analysis with fluorescent probes revealed that ROS levels were elevated within both bacterial cells [20] and fibroblasts [21] after treatment with sub-lethal concentrations of QUATs. In the present work, we used flow cytometric analysis to determine the production of ROS in bacterial cells following exposure to UV-A in both the absence and presence of the geminiQUAT. For hydrogen peroxide generation, losses of cell viability correlated with changes in the induction of hydrogen peroxide production as evidenced by increased fluorescence. These results suggested that the bactericidal action of the combined methods primarily reflects oxidative reaction and damage at several cellular sites, as described for UV-A irradiation in a previous paper [3,6]. The proposed bactericidal role of hydrogen peroxide also agrees with the attenuation of inhibition observed in the presence of catalase. Overproduction of hydrogen peroxide was observed even with combination treatments that used reduced levels of UV-A exposure. Our interpretation is that 3DOBP-4,10, as a geminiQUAT, could be essential for the rapid generation of hydrogen peroxide even with exposure to low fluence dose. Generally, it has been known that superoxide facilitates the production of hydrogen peroxide. To examine possible links between these ROS in this system, the generation of superoxide was investigated by the use of a specific probe. As expected, superoxide fluorescence levels increased significantly in suspensions subjected to combination treatment, when compared with the control sample (30 min incubation). On the other hand, non-significant changes in superoxide levels were detected following exposure to UV-A light alone, demonstrating that UV-A light itself was not sufficient to induce excessive production of superoxide. The results suggest that the overproduced superoxide may instead trigger the rapid production of excessive hydrogen peroxide levels in the combined system, leading to accelerated bactericidal activity. 3DOBP-4,10 was confirmed as an important element for the ROS-overproducing mechanism, because ROS overproduction was attenuated by the addition of albumin as an organic substrate. Other work has demonstrated that organic materials (like albumin) inhibit the function

232

A. Shirai et al. / Journal of Photochemistry and Photobiology B: Biology 130 (2014) 226–233

of dissolved gemini-QUATs that have high hydrophobicity, leading to attenuation of growth inhibition by gemini-QUATs [17]. It has been reported that gemini-QUATs such as 3DOBP-4,10 interact with the bacterial surface and cause (at concentrations of 0.2 lg/ml or less) leakage of small molecules such as magnesium ions and ATP, and respiratory inhibition by action on enzymes located at the cellular membrane [25]. The antimicrobial activity of 3DOBP4,10 appears to be suppressed by depletion of ROS [19]. Therefore, we hypothesize that the synergistic activity of the combination treatment reflects increased oxygen permeation of cells following 3DOBP-4,10-mediated disturbance of the bacterial membrane. In previous work, photo-catalytic decomposition of an unsaturated fatty acid ester, retinyl palmitate, by UV-A irradiation was demonstrated using ESR analysis, proving that singlet oxygen and superoxide were generated in a time-dependent manner during UV-A exposure [26]. It was speculated that the two ROS were formed by the transfer of an electron or energy from the UV-A-excited retinyl palmitate to molecular oxygen. In the combined system, the reaction between the UV-A light-excited lipid membrane and the abundant permeation of oxygen as an acceptor molecule could facilitate ROS generation within bacterial cells, and any superoxide produced would be converted into stronger oxidants such as hydrogen peroxide and hydroxyl radical. Some interference by oxidative reactions may incur extensive damage to bacteria. Hydroxyl radical, the most reactive ROS, is hypothesized to initiate lipid peroxidation, given that the first ROS generated probably is located either within the cell membrane or at a nearby site where the permeation of oxygen and UV-A absorption both occur. The TBARS assay quantitatively detects increased peroxidation by hydroxyl radical [23]. In our combined system, the extent of peroxidation was large. The degree was high even at reduced irradiation doses (2.87 J/cm2 rather than 17.2 J/cm2 of UV-A). Presumably, the large extent of peroxidation was caused by the rapid generation of ROS in the cell membrane, initiated by synergism between 3DOBP-4,10 and UV-A. The bactericidal mechanism of the combination treatment therefore would consist of lethal oxidative damage similar to that seen with UV-A light alone at a higher fluence rate [3]. The combination of biocide with UV-A irradiation at 17.2 J/cm2 induced peroxidation to 8.19 nM MDA equivalents, reducing viability by 4.20-logs (data not shown). The detected peroxidation level was not significantly different from that seen at 2.87 J/cm2. Subsequent to the actions of the gemini-QUAT and the irradiation at the membrane, the oxidative reactions are expected to proceed rapidly, with the oxidative potential achieving saturation in the lipid peroxidation assay system used here. Therefore, we investigated not only peroxidation but also other bactericidal roles of ROS. 8-OHdG is an indicator of DNA oxidation, as expected to result from UV-A exposure [3]. Our combined disinfection procedure led to increases in 8-OHdG concentration, with 8-OHdG levels rising with increasing total fluence dose (from 2.87 to 17.2 J/cm2) and increasing cell killing. At 2.87 J/cm2 UV-A exposure in the combined disinfection procedure, the level of 8-OHdG did not differ significantly compared with the untreated sample (60 min incubation), suggesting that the oxidative agents initiate killing by modifying lipid membranes and proteins present in the membrane. This interpretation is consistent with the finding that the bacterial membrane underwent rapid peroxidation following our combination treatment. Following the initial damage to the membrane, oxidation would proceed to sites deeper within the cells, extending to modification of the bacterial DNA. It has been reported recently that UV-A light can inactivate pathogenic bacteria in water systems [3,4,13]. In order to reduce bacteria by irradiation with a UV-A–LED, total fluence dose at 315 J/cm2 (fluence rate, 70 mW/cm2) was required [3]. Similar results were reported using a treatment that combined hydrogen

peroxide with a higher fluence rate (110 mW/cm2) light source [13]. On the other hand, our disinfectant system permitted the use of a UV-A–LED source with low fluence rate (4.77 mW/cm2). At this low fluence rate, the light source alone was not effective in killing bacterial cells, but the combination of such a light source with a gemini-QUAT such as 3DOBP-4,10 provided synergistic increases in bacterial killing. The combination exhibited strong bactericidal activity against pathogens, even at concentrations of the gemini-QUAT that were not bactericidal on their own. Optimization of the results presented here may provide novel treatments for disinfection of contaminated liquids or of articles intended for use in patient care.

References [1] I.L. Shechmeister, Sterilization by ultraviolet irradiation, in: S.S. Block (Ed.), Disinfection, Sterilization and Preservation, fourth ed., Lea & Febiger, Philadelphia, 1991, pp. 553–565. [2] G.F. Kramer, B.N. Ames, Oxidative mechanisms of toxicity of low-intensity near-UV light in Salmonella typhimurium, J. Bacteriol. 169 (1987) 2259–2266. [3] A. Hamamoto, M. Mori, A. Takahashi, M. Nakano, N. Wakikawa, M. Akutagawa, T. Ikehara, Y. Nakaya, Y. Kinouchi, New water disinfection system using UVA light-emitting diodes, J. Appl. Microbiol. 103 (2007) 2291–2298. [4] M. Mori, A. Hamamoto, A. Takahashi, M. Nakano, N. Wakikawa, S. Tachibana, T. Ikehara, Y. Nakaya, M. Akutagawa, Y. Kinouchi, Development of a new water sterilization device with a 365 nm UV–LED, Med. Biol. Eng. Comput. 45 (2007) 1237–1241. [5] M. Gadelmoula, X. Lian, M. Maeda, M. Aihara, K. Mawatari, A. Hamamoto, Y. Harada, M. Yamato, M. Akutagawa, Y. Nakaya, Y. Kinouchi, A. Takahashi, Suitability of ultraviolet (A)–light emitting diode for air stream disinfection, J. Med. Invest. 56 (2009) 150–156. [6] J. Chamberlain, S.H. Moss, Lipid peroxidation and other membrane damage produced in Escherichia coli K1060 by near-UV radiation and deuterium oxide, Photochem. Photobiol. 45 (1987) 625–630. [7] M.J. Davies, Singlet oxygen-mediated damage to proteins and its consequences, Biochem. Biophys. Res. Commun. 305 (2003) 761–770. [8] M. Pezzoni, R.A. Pizarro, C.S. Costa, Protective effect of low UVA irradiation against the action of lethal UVA on Pseudomonas aeruginosa: role of the relA gene, J. Photochem. Photobiol., B 116 (2012) 95–104. [9] S. Arikan, M. Lozano-Chiu, V. Paetznick, J.H. Rex, In vitro synergy of caspofungin and amphotericin B against Aspergillus and Fusarium spp., Antimicrob. Agents Chemother. 46 (2002) 245–247. [10] P.-C. Maness, S. Smolinski, D.M. Blake, Z. Huang, E.J. Wolfum, W.A. Jacoby, Bactericidal activity of photocatalytic TiO2 reaction: toward an understanding of its killing mechanism, Appl. Environ. Microbiol. 65 (1999) 4094–4098. [11] Z. Huang, P.-C. Maness, D.M. Blake, E.J. Wolfrum, S.L. Smolinski, W.A. Jacoby, Bactericidal mode of titanium dioxide photocatalysis, J. Photochem. Photobiol. A: Chem. 130 (2000) 163–170. [12] N.T.T. Le, H. Nagata, M. Aihara, A. Takahashi, T. Okamoto, T. Shimohata, K. Mawatari, Y. Kinouchi, M. Akutagawa, M. Haraguchi, Additional effects of silver nanoparticles on bactericidal efficiency depend on calcination temperature and dip-coating speed, Appl. Environ. Microbiol. 77 (2011) 5629–5634. [13] M. Shirato, H. Ikai, K. Nakamura, E. Hayashi, T. Kanno, K. Sasaki, M. Kohno, Y. Niwano, Synergistic effect of thermal energy on bactericidal action of photolysis of H2O2 in relation to acceleration of hydroxyl radical generation, Antimicrob. Agents Chemother. 56 (2012) 295–301. [14] Y.A. Khan, R.T. Kashiwabuchi, S.A. Martins, J.M. Castro-Combs, S. Kalyani, P. Stanley, D. Flikier, A. Behrens, Riboflavin and ultraviolet light A therapy as an adjuvant treatment for medically refractive Acanthamoeba keratitis, Ophthalmology 118 (2011) 324–331. [15] M.J. Rosen, J.H. Mathias, L. Davenport, Aberrant aggregation behavior in cationic gemini surfactants investigated by surface tension, interfacial tension, and fluorescence methods, Langmuir 15 (1999) 7340–7346. [16] A. Shirai, T. Maeda, H. Nagamune, H. Matsuki, S. Kaneshina, H. Kourai, Biological and physicochemical properties of gemini quaternary ammonium compounds in which the positions of a cross-linking sulfur in the spacer differ, Eur. J. Med. Chem. 40 (2005) 113–123. [17] H. Kourai, T. Yabuhara, A. Shirai, T. Maeda, H. Nagamune, Syntheses and antimicrobial activities of a series of new bis-quaternary ammonium compounds, Eur. J. Med. Chem. 41 (2006) 437–444. [18] A. Shirai, T. Sumitomo, M. Yoshida, T. Kaimura, H. Nagamune, T. Maeda, H. Kourai, Synthesis and biological properties of gemini quaternary ammonium compounds, 5,50 -[2,20 -(a,x-polymethylnedicarbonyldioxy)diethyl]bis(3alkyl-4-methylthiazolium iodide) and 5,50 -[2,20 -(pphenylenedicarbonyldioxy)diethyl]bis(3-alkyl-4-methylthiazolium bromide), Chem. Pharm. Bull. 54 (2006) 639–645. [19] A. Shirai, S. Ueta, H. Maseda, H. Kourai, T. Omasa, Action of reactive oxygen species in the antifungal mechanism of gemini-pyridinium salts against yeasts, Biocontrol Sci. 17 (2012) 77–82.

A. Shirai et al. / Journal of Photochemistry and Photobiology B: Biology 130 (2014) 226–233 [20] K. Nakata, T. Tsuchido, Y. Matsumura, Antimicrobial cationic surfactant, cetyltrimethylammonium bromide, induces superoxide stress in Escherichia coli cells, J. Appl. Microbiol. 110 (2010) 568–579. [21] S. Ma, L. Shan, Y. Xiao, F. Li, H. Li, S. Lijuan, J. Chen, The cytotoxicity of methacryloxylethyl cetyl ammonium chloride, a cationic antibacterial monomer, is related to oxidative stress and the intrinsic mitochondrial apoptotic pathway, Braz. J. Med. Biol. Res. 44 (2011) 1125–1133. [22] T. Dias, B. Liu, P. Jones, P.J. Houghton, H. Mota-Filipe, A. Paulo, Cytoprotective effect of Coreopsis tinctoria extracts and flavonoids on tBHP and cytokineinduced cell injury in pancreatic MIN6 cells, J. Ethnopharmacol. 139 (2012) 485–492. [23] K. Nakamura, Y. Yamada, H. Ikai, T. Kanno, K. Sasaki, Bactericidal action of photoirradiated gallic acid via reactive oxygen species formation, J. Agric. Food Chem. 60 (2012) 10048–10054.

233

[24] S.A.R. Martins, J.C. Combs, G. Noguera, W. Camacho, P. Wittmann, R. Walther, M. Cano, J. Dick, A. Behrens, Antimicrobial efficacy of riboflavin/UVA combination (365 nm) in vitro for bacterial and fungal isolates: a potential new treatment for infectious keratitis, Invest. Ophthalmol. Vis. Sci. 49 (2008) 3402–3408. [25] T. Sumitomo, T. Maeda, H. Nagamune, H. Kourai, Bacterioclastic action of a bisquaternary ammonium compound against Escherichia coli, Biocontrol Sci. 9 (2004) 1–9. [26] Q. Xia, J.J. Yin, S.-H. Cherng, W.G. Wamer, M. Boudreau, P.C. Howard, P.P. Fu, UVA photoirradiation of retinyl palmitate-formation of singlet oxygen and superoxide, and their role in induction of lipid peroxidation, Toxicol. Lett. 163 (2006) 30–43.