New insights into the aquatic photochemistry of fluoroquinolone antibiotics: Direct photodegradation, hydroxyl-radical oxidation, and antibacterial activity changes

Highlights

• It is first reported on hydroxyl-radical oxidation of 6 fluoroquinolone antibiotics.

• Methods were developed to assess photolysis and oxidation fate in surface waters.

• The neutral form reacted faster with hydroxyl radical than protonated forms.

• The main oxidation intermediates and transformation pathways were clarified.

• The antibacterial activity changes depend on dominant photolysis pathways.

Abstract

The ubiquity and photoreactivity of fluoroquinolone antibiotics (FQs) in surface waters urge new insights into their aqueous photochemical behavior. This study concerns the photochemistry of 6 FQs: ciprofloxacin, danofloxacin, levofloxacin, sarafloxacin, difloxacin and enrofloxacin. Methods were developed to calculate their solar direct photodegradation half-lives (t_d,E) and hydroxyl-radical oxidation half-lives (t_OH,E) in sunlit surface waters. The t_d,E values range from 0.56 min to 28.8 min at 45° N latitude, whereas t_OH,E ranges from 3.24 h to 33.6 h, suggesting that most FQs tend to undergo fast direct photolysis rather than hydroxyl-radical oxidation in surface waters. However, a case study for levofloxacin and sarafloxacin indicated that the hydroxyl-radical oxidation induced risky photochlorination and resulted in multi-degradation pathways, such as piperazinyl hydroxylation and clearage. Changes in the antibacterial activity of FQs caused by photodegradation in various waters were further examined using Escherichia coli, and it was found that the activity evolution depended on primary photodegradation pathways and products. Primary intermediates with intact FQ nuclei retained significant antibacterial activity. These results are important for assessing the fate and risk of FQs in surface waters.

Introduction

Fluoroquinolone antibiotics (FQs) are important aqueous micropollutants that are not effectively removed by wastewater treatment plants (Chen et al., 2013) and are confirmed to be ubiquitous in surface waters (Schwarzenbach et al., 2006, Luo et al., 2011, Chen and Zhou, 2014, Liang et al., 2013). This large class of pollutants can promote bacterial resistance and photomodified toxicities, making them an emerging concern (Ge et al., 2010, Li et al., 2011). In aquatic systems, FQs show resistance to hydrolysis and biodegradation (Ge et al., 2010, Knapp et al., 2005). However, exposure to sunlight will weaken the persistence of FQs in the aqueous euphotic zone (Ge et al., 2010). Thus, new insights into their environmental photochemical behavior and associated risk are required to assess the fate and effects of FQs in sunlit surface waters.

FQs can undergo not only direct or indirect photodegradation but also self-sensitized photo-oxidation via reactive oxygen species (ROS, such as hydroxyl radicals, OH) (Ge et al., 2010, Li et al., 2011, Sturini et al., 2010, Wei et al., 2013). Many other antibiotics, such as tetracycline (Werner et al., 2006, Chen et al., 2008), and phenicols (Ge et al., 2009), are also self-sensitized photodegradable. Werner et al. (2006) investigated the photochemical kinetics of tetracycline and found that the pseudo-first-order rate constants (k) increased with the increasing initial concentration (C₀), likely because of the self-sensitization effect. Because FQs can self-sensitize, their photolytic k might be affected by C₀. To verify the hypothesis, it is necessary to examine the C₀-affected kinetics and to weigh the contribution of self-sensitized photolysis to total FQ elimination.

Many previous studies have calculated the environmental direct photolytic half-lives (t_d,E) for various micropollutants in surface waters based on the method of Leifer (1988) (Boreen et al., 2004, Boreen et al., 2005, Edhlund et al., 2006). However, the FQ t_d,E values reported by different investigators differ at orders of magnitude. For example, Burhenne et al. (1997) calculated t_d,E = 1.6 h (summer)–55.4 h (winter) for the direct photolysis of enrofloxacin (ENR) in surface waters at 30°–60° N latitude, whereas Ge et al. (2010) reported the t_d,E = 1.25 min (summer)–6.11 min (winter) at 45° N latitude. Therefore, it is necessary to examine the distinct difference of t_d,E, and develop a calculation method for evaluating t_d,E might be beneficial to better understand this behavior.

Some previous studies that reported the phototransformation pathways of FQs in water mainly focused on their direct photolysis (Ge et al., 2010, Sturini et al., 2010, Babić et al., 2013, Sturini et al., 2012a). Ge et al. (2010) found that photoinduced decarboxylation, defluorination, or piperazinyl N⁴-dealkylation occurred effectively depending on the structures of the FQs, whereas Sturini et al. (2010) identified a completely different path, i.e., photopolymerization for marbofloxacin. Additionally, FQs might undergo indirect photodegradation or sensitized photooxidation mediated by the photogenerated OH from dissolved organic matter, Fe(III) and nitrate (Mack and Bolton, 1999, Grannas et al., 2014, Fisher et al., 2006, Walse et al., 2004). The hydroxyl-radical oxidation is also common in advanced oxidation (Keen and Linden, 2013, DeWitte et al., 2008) and self-sensitization processes (Ge et al., 2010), which indicate that the bimolecular reactions between OH and FQs might be the relevant mechanisms in these processes. However, information regarding the OH-initiated kinetics and associated photoproducts of FQs remains limited.

Because OH is ubiquitous in sunlit surface waters and shows less selectivity to oxidize most organic chemicals (Keen and Linden, 2013, Mill, 1999, Li et al., 2014), the OH oxidation reaction of FQs may result in more degradation pathways than direct photolysis. Moreover, the different FQ protonation states are supposed to react with OH at different reactivities. Therefore, the environmentally apparent OH-oxidation half-lives (t_OH,E) for FQs in surface water will depend on the ambient pH value. Previously, photooxidation reactivities toward OH were reported for each protonated state of a few chemicals, such as hydroxylated polybrominated diphenyl ethers (Xie et al., 2013). For antibiotics, they can react with OH at mean rate constants of 5.3 × 10⁹ M^− 1 s^− 1, 5.0 × 10⁹ M^− 1 s^− 1 and 3.8 × 10⁹ M^− 1 s^− 1 for sulfa drugs (Boreen et al., 2004, Boreen et al., 2005), nitrofurans (Edhlund et al., 2006) and ionophores (Yao et al., 2013), respectively. However, their reactivities toward OH of different protonated states are still largely unknown.

Because photodegradation is an important transformation pathway for FQs in surface waters, it is of great significance to investigate their antibacterial activity changes caused by photodegradation. For ciprofloxacin (CIP), Paul et al. (2010) found that the antibacterial potency to Escherichia coli (E. coli) of irradiated solutions decreased and that the photolytic products retained negligible antibacterial activity relative to the parent compound. However, Sturini et al. (2012b) showed notable antimicrobial activity of E. coli to the ciprofloxacin photoproducts. Furthermore, norfloxacin and ofloxacin photoproducts were found to be inactive to E. coli, whereas ENR photoproducts retained significant activity (Wammer et al., 2013). These results appear to be inconsistent, suggesting that antibacterial activity changes of FQs because of their photodegradation need to be further explored. For other classes of antibiotics, such as tetracycline, triclosan and sulfa drugs, changes in antibacterial activity caused by photochemical transformations have been studied previously (Wammer et al., 2011, Wammer et al., 2006).

This study seeks new insights into the aquatic photochemical behavior of 6 FQs: CIP, danofloxacin (DAN), levofloxacin (LEV), sarafloxacin (SAR), difloxacin (DIF), and ENR. The objectives are to develop methods to calculate the environmental half-lives (t_d,E and t_OH,E), investigate OH-oxidation kinetics and pathways of selected FQs, and examine changes in antibacterial activity because of photodegradation. To the best of our knowledge, this is the first comprehensive report on the kinetics and mechanisms for OH-mediated photodegradation of FQs.