The contamination of triclosan, which is a widely used antimicrobial agent, is of emerging concern for its potential toxicity to aquatic organisms and human beings. Chlorination, UV254 irradiation and ozonation are the main disinfection methods used in water treatment processes. Although studies have shown that triclosan could react with free chlorine and ozone, and undergo photolysis under UV irradiation, few of them focused on the effects of disinfection processes on the biotoxicity of triclosan. In the present study, the potential risk of triclosan in water before and after laboratory-scale disinfection processes, including chlorination, UV254 irradiation and ozonation, was evaluated by bioassay. The results showed that both acute toxicity and genetic toxicity of triclosan was increased by chlorination and UV254 disinfection but decreased by ozone disinfection. In other words, ozonation would be a preferential disinfection method for triclosan-containing surface waters. This finding will help us to choose an appropriate disinfection method for water treatment. Furthermore, it is proved that bioassay could be a feasible way for risk evaluation when concentrations of antibiotics in surface waters or drinking waters are very low.
Triclosan (TCS) is a broad-spectrum antimicrobial agent widely used in personal care products such as soap, toothpaste, shampoo, disinfectant, skin cream, and so on (Bester 2005). The large amount of triclosan from daily use enters municipal wastewater treatment plants together with the sewage and eventually is discharged with wastewater effluent. Triclosan is frequently detected in surface waters. The reported highest concentration of triclosan reached 2.3 μg/L in surface waters and 15 ng/L in drinking water (Bedoux et al. 2012; Dhillon et al. 2015). The contamination of triclosan has generated great public health concerns because of its potential toxicity toward aquatic organisms and human beings. It was reported that triclosan-contaminated wastewaters could affect the structure and function of algal communities (Reiss et al. 2002). By in vitro and in vivo studies, it has been proved that triclosan could present acute toxicity to fish, crustaceans, microalgal species, higher plants and insects (Franz et al. 2008; Bedoux et al. 2012; Dhillon et al. 2015). Triclosan also showed the cytotoxic and genotoxic effects to zebra mussels, zebrafish and sea urchin (Binelli et al. 2009; Oliveira et al. 2009; Hwang et al. 2014). Besides, triclosan was observed to affect the metabolism of thyroidal hormones of mice, decrease the sperm production in male rats and cause negative effects in the fetus development of sheep placenta, etc., which might be due to its similar chemical structure to some estrogens (Matsumura et al. 2005; Kumar et al. 2009).
Chlorination, UV254 irradiation and ozonation are the main disinfection methods used in water treatment processes. Sodium hypochlorite is a source of free chlorine commonly used as a disinfecting oxidant. Studies have shown that triclosan could be chlorinated by free chlorine to form byproducts including chlorinated phenoxy-phenols, chlorinated phenols, and trihalomethanes (Rule et al. 2005; Fiss et al. 2007). UV254 disinfection has potential to reduce the production of disinfection byproducts and is receiving more and more attention in water supply treatment. However, some of the residual organic pollutants in surface waters are readily to undergo photolysis under UV irradiation. It has been reported that triclosan could absorb UV light and undergo direct photolysis (Tixier et al. 2002; Chen et al. 2016b). A major concern on the photochemical degradation of triclosan is the potential formation of highly toxic and persistent dioxin products (e.g. 2,8-dichlorodibenzo-p-dioxin), which might be formed via an intramolecular photochemical substitution reaction under UV irradiation (Mezcua et al. 2004; Latch et al. 2005). Ozone is a very powerful disinfecting and deodorizing gas, widely used in removing bacteria, viruses, algae and fungi as well as oxidizing and mineralizing organic chemicals (Wert et al. 2009).
Although studies have shown that triclosan could react with the disinfectants or undergo photolysis during the disinfection processes, few of them focused on the changes in biotoxicity of triclosan after disinfection. Because of the low concentration of triclosan in the surface water and drinking water, it is difficult to evaluate the risk by identifying the products and determining their concentrations. Thus, the objective of the present study was to evaluate the potential risk of triclosan in water before and after the conventional disinfection processes, including chlorination, UV254 irradiation and ozonation by bioassay.
MATERIALS AND METHODS
Triclosan (powder, ≥97%), were purchased from Sigma-Aldrich (St Louis, MO, USA). All chemicals and solvents used were of purity of at least analytical-reagent grade. Ultrapure water (electrical resistivity >18.2 MΩcm) was provided by a Millipore purification system (Millipore S.A.S, Molsheim, France). The stock solution of triclosan with a concentration of 100 mg/L was prepared in acetonitrile (HPLC grade, Fisher Chemical, USA). Each solution for the disinfection experiment contained 0.5 mg/L of triclosan in the ultrapure water and the initial pH value of the triclosan solution was 6.8.
Various doses for each disinfectant (chlorine, UV, and ozone) were used in this study, including a low dose which was close to the practical dose used in water treatment plants and another two higher doses in order to confirm the influence of disinfection process on the biotoxicity according to the literature. The chlorination of triclosan solutions was carried out in 250-mL Erlenmeyer flasks with magnetic stir bars to mix samples gently at 25 °C. Sodium hypochlorite was added to achieve the initial various doses of chlorination (5, 10 and 20 mg Cl2/L) for a contact time of 30 min (Buth et al. 2011; Zhang et al. 2015). Then sodium thiosulfate solution (0.5%) was added to the remainder to terminate chlorination process. Samples (10 mL) were collected from each solution before and after chlorination process for the bioassay. The pH of the sample after chlorination was adjusted to 7.0 by adding hydrochloric acid and sodium hydroxide before the toxicity tests. For laboratory-scale UV254 disinfection, 50-mL quartz glass tubes with triclosan solutions inside were placed in a photochemical reactor (YM-GHX-V, Yuming Instrument Company, China) with a low-pressure mercury vapor 254 nm lamp (TUV- T54P-SE, Philips) in the center. Each solution was stirred gently by a magnetic stir bar. The UV lamp was pre-warmed for at least 15 min to ensure stable UV irradiation fluence. The fluence rate was fixed at 0.28 mW/cm2 determined by an ultraviolet radiometer (UV-B, Handy, China). Samples (10 mL) were withdrawn after 0, 1, 2, and 5 min of irradiation for the bioassay, and the corresponding UV doses were 0, 16.8, 33.6, and 84.0 mJ/cm2 (Zhang et al. 2015; Ben et al. 2016). Ozone disinfection was performed in a 500 mL-glass bottle. Ozone generated from pure oxygen by an ozone generator (VMUS-4, AZCO, Canada) was introduced into the glass bottle and the concentration of ozone gas was measured by an ozone analyzer (UV100, ECO, USA). The triclosan solutions were treated by different concentrations of ozone gas (0.5, 1, 2 mg/L) for a contact time of 5 min (Chen et al. 2012; Orhon et al. 2017). The pH values of the triclosan solutions were determined during UV and ozone disinfection, and the results showed that the pH values were almost unchanged during the disinfection. All the disinfection processes were conducted in duplicate at 25 °C.
The acute toxicity of each triclosan solution before and after disinfection was determined by the bioluminescence inhibition assay (Villa et al. 2014; Chen et al. 2016a). The freeze-dried bacteria Vibrio qinghaiensis sp. nov. (Q67) obtained from East China Normal University was first revived before the test. 0.1 mL of cultivated bacterial suspension and 0.9 mL of each test solution were mixed in a quartz tube. Ultrapure water was used as the blank control. The relative light unit of each test medium and the blank control were detected by a luminator (Berthold, Germany) after an exposure of 20 min. The acute toxicity was thus evaluated by the relative inhibition of light emission and the toxicity was calculated by comparison between the relative light unit of each test medium and the blank control. The genetic toxicity of each triclosan solution before and after disinfection was tested by the Vicia faba roots tip micronucleus assay (Chen et al. 2016a). Dry broad bean (Vicia faba) seeds were purchased from Central China Normal University, which were harvested from the plants grown in an unpolluted area. The beans were firstly soaked and germinated at 25 °C until 2–3 cm roots emerged. Five germinated seeds with homogeneous growth roots were exposed to each test solution and the blank control (ultrapure water) for 6 h before cultivated at 25 °C for 22 h. Then the root tips of the beans were cut off and fixed by Carnoy liquid (acetic acid/ethanol, 1/3) before stained by the Feulgen staining technique. Finally, the root tips were squashed on slides and the amount of micronuclei on each slide was counted under a microscope. The genetic toxicity was evaluated by the frequency of micronuclei which was the average number of micronuclei per 1,000 cells for each root tip.
All statistical analyses were conducted with Statistical Package for the Social Sciences 16.0 for Windows. One-way analysis of variance (ANOVA) was applied to detect significant differences and the statistical significance was accepted at p < 0.05.
RESULTS AND DISCUSSION
Compared to the blank control, the triclosan solution before disinfection displayed significant acute and genetic toxicity, with the luminescence inhibition percentage of 11.32% and micronucleus frequency of 7.33‰, as shown in Figure 1. Triclosan has been reported to pose acute toxicity and genotoxic effects to various testing organisms, including crustaceans, zebra mussels, zebrafish, microalgal, insects and rats (Binelli et al. 2009; Bedoux et al. 2012; Hwang et al. 2014; Dhillon et al. 2015). Even at a very low concentration, triclosan in the surface water and drinking water still presents potential health risks.
The comparison of the acute and genetic toxicity of triclosan solutions before and after chlorination is shown in Figure 1. The chlorination process increased both acute and genetic toxicity of the solutions significantly. Moreover, with the increase of the chlorination concentration, the acute and genetic toxicity increased accordingly. When the chlorination concentration reached 20 mg/L, the acute toxicity and genetic toxicity of the triclosan solution were almost twice of those before chlorination. It has been reported that triclosan can react quickly with free chlorine to produce phenols, 2,4-dichlorophenol, 2,4,6-trichlorophenol, and chloroform in water disinfection processes (Rule et al. 2005; Fiss et al. 2007). Chloroform, a well-known disinfection byproduct in the chlorine disinfection process, has been classified as a probable human carcinogen by the U.S. Environmental Protection Agency (USEPA). Besides, three chlorinated triclosan derivative products including 4,5-dichloro-2-(2,4-dichlorophenoxy)phenol, 5,6-dichloro-2-(2,4-dichlorophenoxy)phenol, and 4,5,6-trichloro-2-(2,4-dichlorophenoxy)phenol were detected at a level of 22 ng/L in the final treatment plant effluent after chlorine disinfection (Buth et al. 2011). According to Figure 1, it is suggested that these chlorinated triclosan derivatives and disinfection byproducts showed higher acute and genetic toxicity than triclosan. Similar findings were also observed in the study of acetaminophen and 5,5-diphenylhydantoin for more toxic chlorinated byproducts formed during the chlorination process (Mansor & Tay 2017; Ding et al. 2018).
The acute and genetic toxicity of triclosan solutions before and after UV254 irradiation is shown in Figure 2. Similarly, a significant increase in both acute and genetic toxicity of triclosan solution was observed. With the increase of UV254 irradiation fluence, the toxicity increased accordingly. When the UV254 irradiation fluence was increased to 84 mJ/cm2, the luminescence inhibition percentage reached 18.33% and micronucleus frequency reached 12.56‰. Triclosan can undergo direct photolysis under UV irradiation, and monochlorophenol, 2,4-dichlorophenol, 2,4,6-trichlorophenol dichlorohydroxydiphenyl ether, polychlorodibenzo-p-dioxins and polymerized products have been reported as the main direct photolysis products of triclosan (Sanchez-Prado et al. 2006; Chen et al. 2016b). Among these products, polychlorodibenzo-p-dioxins could show very high acute and genetic toxicity, whose formation yield was nearly 4% during the direct photolysis process (Kliegman et al. 2013). Nevertheless, chlorophenols have been proven to show lower acute toxicity and chronic toxicity to fish, daphnid and green algae than triclosan (Gao et al. 2014). Overall, the increased acute toxicity and genetic toxicity indicates the formation of more toxic photo-transformation products of triclosan under UV254 irradiation. However, in the study of 5,5-diphenylhydantoin, the biotoxicity of the solution decreased after UV disinfection, suggesting that the products were less toxic (Mansor & Tay 2017).
However, different results were observed in ozone disinfection. As shown in Figure 3, slight influence was shown by ozone disinfection when the ozone concentration was as low as 0.5 mg/L; but significant decrease in both acute toxicity and genetic toxicity was observed when the ozone concentration was higher. Ozonation, which has been recently proposed as an important advanced oxidation technology for the removal of a wide range of organic pollutants in wastewater, was proven to be an effective post-treatment technique for pharmaceuticals and personal care products (Wert et al. 2009). The removal of triclosan by aqueous ozone has been studied and the main transformation products were identified as 2,4-dichlorophenol, 4-chlorocatecol, mono-hydroxy-triclosan and di-hydroxy-triclosan during the treatment process (Orhon et al. 2017). When the ozone concentration was high enough, organic compounds would even be mineralized (Wert et al. 2009). Although 2,4-dichlorophenol was proved to be harmful to organisms, the results of bioassay suggested lower risk of triclosan solutions after ozone disinfection. In the study of oxcarbazepine, significant increase in the toxicity by Vibrio fischeri bioluminescence was observed (Wang et al. 2018). However, when the dose of ozone was high enough, the toxicity of organic compound was reduced due to the mineralization (Lovato et al. 2017).
Antibiotics including triclosan are widely used, but they could not be totally removed by conventional processes in municipal wastewater treatment plants, resulting in the inevitable presence of residual antibiotics in surface waters. Many of them undergo transformation during the conventional disinfection processes such as chlorination, UV254 irradiation and ozonation. Consequently, more attention should be paid to the risk caused by disinfection processes. As shown in our study, both acute toxicity and genetic toxicity of triclosan was increased by chlorination and UV254 disinfection but decreased by ozone disinfection. This might inspire us to choose an appropriate disinfection method. Besides, our results proved that bioassay could be a feasible way for risk evaluation when the surface water or drinking water contains very low concentrations of antibiotics.
This work was supported by the National Natural Science Foundation of China (No. 41301545) and the Priority Academic Program Development of Jiangsu Higher Education Institutions Fund. The authors also want to thank Prof. Zhi Zhou from Purdue University for language checking this paper.