Volatile DBPs contributed marginally to the developmental toxicity of drinking water DBP mixtures against Platynereis dumerilii

Highlights

• Concentrated water samples were prepared with high concentrations of NOM, bromide and chlorine.

• Nitrogen sparging removed 98.5–99.7% of volatile DBPs in the concentrated water samples.

• No significant difference in developmental toxicity was found for each sample without and with N₂ sparging.

• The contribution of volatile DBPs to the developmental toxicity of DBP mixtures was marginal.

Abstract

Halogenated disinfection byproducts (DBPs) are formed during chlorine disinfection of drinking water. The complicated natural organic matter in source water causes the formation of an even more complicated mixture of DBPs. To evaluate the toxicity of a DBP mixture in a disinfected water sample, the sample needs to be pretreated in order to attain an observable acute adverse effect in the toxicity test. During sample pretreatment, volatile DBPs including trihalomethanes, haloacetonitriles and haloketones may be lost, which could affect the toxicity evaluation of the DBP mixture. In this study, we intentionally prepared “concentrated” simulated drinking water samples, which contained sufficiently high levels of volatile and nonvolatile DBPs and thus enabled directly evaluating the toxicity of the DBP mixtures without sample pretreatment. Specifically, the natural organic matter and bromide concentrations and the chlorine dose in the concentrated water samples were 250 times higher than those in a typical drinking water sample. Each concentrated water sample was divided into two aliquots, and one of them was nitrogen sparged to eliminate volatile DBPs; then, both aliquots were used directly in a well-established developmental toxicity test. No significant difference (p > 0.10) was found between the developmental toxicity indexes of each concentrated water sample without and with nitrogen sparging, indicating that the contribution of volatile DBPs to the developmental toxicity of the DBP mixture might be marginal. A reasonable interpretation is that nonvolatile halogenated DBPs (especially the aromatic ones) in the DBP mixture could be the major developmental toxicity contributor that warrants more attention.

Introduction

Water disinfection achieves great success in inactivating harmful pathogens and preventing acute waterborne diseases. Chlorination is the most extensively used technology due to its convenience, low cost and high effectiveness in controlling pathogenic microbes (Li et al., 2017a). However, the reactions of chlorine with natural organic matter (NOM) and inorganic ions (e.g., bromide) in source water unintendedly generate halogenated disinfection byproducts (DBPs) (Richardson et al., 2007; Xie, 2003; Hu et al., 2018; He et al., 2018; Wu et al., 2017). Since the discovery of trichloromethane (TCM) as the first DBP in chlorinated water in 1974, over 700 halogenated DBPs have been reported in the literature to date (Rook, 1974; Richardson and Ternes, 2017), and they contributed to less than 50% of the total halogenated DBPs formed in chlorinated water (Richardson et al., 2007). Notably, less than 30 of the identified DBPs are considered as volatile DBPs with the characteristics of boiling points below 250 °C at 760 mmHg and vapor pressures above 2.0 mmHg at 25 °C (WHO, 1989), e.g., trihalomethanes (THMs), haloacetonitriles (HANs), and haloketones (HKs) as shown in Table S1 of Supporting Information (SI). Among the volatile DBPs, six of them are regulated by the World Health Organization (WHO) and the U.S. Environmental Protection Agency (EPA) (Richardson and Postigo, 2012), including four THMs and two HANs: TCM, dibromochloromethane (DBCM), bromodichloromethane (BDCM), tribromomethane (TBM), dichloroacetonitrile (DCAN) and dibromoacetonitrile (DBAN). Epidemiological studies have indicated that human exposure to DBPs was associated with increased risks of bladder and colorectal cancers, asthma and other respiratory complications, potential adverse reproductive and developmental effects (Costet et al., 2011; Villanueva et al., 2015; van Veldhoven et al., 2018). Because of concerns over the harmful effects of these DBPs, WHO has recommended specific maximum values for TCM, DBCM, BDCM, TBM, DCAN, and DBAN in the WHO Guidelines at 200, 100, 60, 100, 20, and 70 μg/L, respectively (Richardson and Postigo, 2012; Richardson, 2005). Besides, U.S. EPA has regulated the maximum contaminant concentration of THMs at 80 μg/L in the Disinfectants/DBP Rule (USEPA, 2006).

Recently, over 100 new halogenated DBPs were detected, identified and quantified in chlorinated waters with mass spectrometry-based approaches (Zhai and Zhang, 2011; Zhao et al., 2012; Li et al., 2012, 2016; Pan and Zhang, 2013; Pan et al., 2016; Han et al., 2017; Zhang et al., 2018; Carter et al., 2019; Huang et al., 2019). Most of the newly identified DBPs in chlorinated drinking waters are aromatic chlorinated and brominated DBPs. Toxicological studies have shown that aromatic halogenated DBPs generally have significantly higher genotoxicity, cytotoxicity, developmental toxicity, and growth inhibition than aliphatic DBPs, and brominated DBPs generally have significantly higher toxicity than their chlorinated analogues (Yang and Zhang, 2013; Liu and Zhang, 2014; Wagner and Plewa, 2017). However, the toxicity of the currently identified DBPs (including regulated and emerging DBPs) in a chlorinated water sample cannot fully reflect the overall toxicity induced by a DBP mixture (Itoh et al., 2011; Zhu and Zhang, 2016). Most of the 700 reported DBPs have not been investigated toxicologically and numerous DBPs remain unidentified. Besides, the potential interactive effects among DBPs on different test organisms can be synergistic, additive, or antagonistic (Han and Zhang, 2018; Rider et al., 2018). Thus, increasing attention has been paid to evaluate the overall toxicity induced by DBP mixtures (Speth et al., 2008; Stalter et al., 2016; Jiang et al., 2017; Abusallout et al., 2017; Han and Zhang, 2018). Notably, various factors such as pH, chlorine contact time, and bromide concentration may affect the formation and speciation of halogenated DBPs during chlorination. It has been reported that the formation of THMs was increased with sample pH and chlorine contact time (Hua and Reckhow, 2008), and increasing the bromide concentration shifted DBP species from being less brominated to being more brominated (Pan and Zhang, 2013; Roccaro et al., 2013). Consequently, the overall toxicity of a chlorinated water sample may be influenced by these factors that affect the formation and shift the distribution of DBP species in a DBP mixture.

As the DBP concentrations in drinking water are generally below the concentrations that can induce an observable acute effect in in vivo and in vitro bioassays, an appropriate pretreatment is required for concentrating DBPs in drinking water samples prior to toxicity evaluation. Liquid-liquid extraction (LLE) and solid-phase extraction (SPE) are the two widely used pretreatment methods to enrich DBPs in drinking water samples (Zheng et al., 2012; Stalter et al., 2016; Köke et al., 2018). Both LLE and SPE are based on the distribution of target compounds (e.g., DBPs) between an organic solvent/sorbent and water in a definite proportion. To mitigate the adverse effect of the organic solvent, the solvent containing DBPs after extraction is completely removed by rotary evaporation or nitrogen sparging, and the nonvolatile fraction of a DBP mixture is further retained in a solvent (e.g., water or dimethyl sulfoxide) that is compatible with bioassays (Han and Zhang, 2018; Köke et al., 2018). Such an organic solvent exchange in the pretreatment inevitably induces the partial or complete loss of volatile DBPs prior to bioassays. Reverse osmosis (RO) and ion exchange with XAD-resin have also been used to obtain DBP-containing concentrates for testing the toxicity of DBP mixtures (Simmons et al., 2002; Speth et al., 2008), but neither RO membrane nor XAD-resin can enrich volatile DBPs. Most recently, a new approach by combining freeze-drying or rotary-evaporation pretreatment with a high salinity tolerant bioassay has been developed for evaluating the comparative toxicity of DBP mixtures from different disinfection scenarios (Han and Zhang, 2018). Compared with LLE, SPE, RO and ion exchange, the pretreatment with freeze-drying or rotary-evaporation can retain nearly all inorganic DBPs (Han and Zhang, 2018), but it cannot retain volatile DBPs either. It has been reported that volatile DBPs are an important group of halogenated DBPs in disinfected waters, especially in chlorinated drinking water. For instance, THMs can account for more than 20% of total organic halogen in chlorinated drinking water (Zhang et al., 2000; Krasner et al., 2006; Hua and Reckhow, 2007; Plewa et al., 2017). Accordingly, there is a critical concern about whether volatile DBPs significantly affect the overall toxicity of DBP mixtures in chlorinated drinking waters.

This study aimed to evaluate the contribution of volatile DBPs to the developmental toxicity of DBP mixtures by comparing the toxicity of chlorinated drinking waters without and with volatile DBPs. Specifically, “concentrated” simulated drinking water samples were prepared by magnifying NOM, alkalinity and bromide concentrations and chlorine dose by 250 times of those in a typical drinking water sample. These samples were also prepared at different pH values, contact times, and bromide concentrations. These samples without or with nitrogen sparging (corresponding to DBPs mixtures with or without volatile DBPs) might contain sufficiently high levels of DBPs that directly induce observable adverse effects on the embryos of a polychaete Platynereis dumerilii. The embryos of the polychaete can tolerate a high salinity from the high concentrations of inorganic ions in concentrated water samples (Han and Zhang, 2018; Li et al., 2019) and have been used successfully for studying the developmental toxicity of various individual DBPs and drinking water DBP mixtures (Yang and Zhang, 2013; Liu et al., 2015; Jiang et al., 2017; Han and Zhang, 2018; Li et al., 2019).