Environmental Toxicology and Chemistry, Vol. 24, No. 2, pp.

423430, 2005 2005 SETAC Printed in the USA 0730-7268/05 $12.00 .00

TOXICITY OF FLUOROQUINOLONE ANTIBIOTICS TO AQUATIC ORGANISMS

APRIL A. ROBINSON, JASON B. BELDEN, and MICHAEL J. LYDY*
Fisheries and Illinois Aquaculture Center and Department of Zoology, Southern Illinois University at Carbondale, Carbondale, Illinois 62901, USA ( Received 26 April 2004; Accepted 2 August 2004) AbstractToxicity tests were performed with seven uoroquinolone antibiotics, ciprooxacin, lomeoxacin, ooxacin, levooxacin, clinaoxacin, enrooxacin, and umequine, on ve aquatic organisms. Overall toxicity values ranged from 7.9 to 23,000 g/L. The cyanobacterium Microcystis aeruginosa was the most sensitive organism (5-d growth and reproduction, effective concentrations [EC50s] ranging from 7.9 to 1,960 g/L and a median of 49 g/L), followed by duckweed (Lemna minor, 7-d reproduction, EC50 values ranged from 53 to 2,470 g/L with a median of 106 g/L) and the green alga Pseudokirchneriella subcapitata (3-d growth and reproduction, EC50 values ranged from 1,100 to 22,700 g/L with a median 7,400 g/L). Results from tests with the crustacean Daphnia magna (48-h survival) and fathead minnow (Pimephales promelas, 7-d early life stage survival and growth) showed limited toxicity with no-observed-effect concentrations at or near 10 mg/L. Fish dry weights obtained in the ciprooxacin, levooxacin, and ooxacin treatments (10 mg/L) were signicantly higher than in control sh. The hazard of adverse effects occurring to the tested organisms in the environment was quantied by using hazard quotients. An estimated environmental concentration of 1 g/L was chosen based on measured environmental concentrations previously reported in surface water; at this level, only M. aeruginosa may be at risk in surface water. However, the selective toxicity of these compounds may have implications for aquatic community structure. KeywordsFluoroquinolone antibiotics Toxicity testing Aquatic organisms Pharmaceuticals

INTRODUCTION

Fluoroquinolones (FQs) are a class of antibiotics that are commonly used in both human and veterinary medicine. These drugs inhibit key bacterial enzymes (DNA gyrase and topoisomerase IV) involved in unwinding the DNA helix for replication and transcription [1]. The structures of FQs used in this study are shown in Figure 1. The rst-generation or original quinolones were not uorinated and included such drugs as nalidixic acid and oxolinic acid. The addition of a uorine atom to the structure enhanced the antimicrobial activity against gram-negative bacteria and extended its activity to some gram-positive bacteria. These antibiotics are known as the second-generation FQs and include such drugs as ciprooxacin (CIP), lomeoxacin (LOM), and ooxacin (OFL). The third-generation FQs, such as levooxacin (LEV), have expanded activity against gram-positive bacteria, while retaining broad gram-negative coverage. Last, the fourth-generation FQs, such as clinaoxacin (CLI), inhibit anaerobic bacteria, while still retaining the gram-negative and gram-positive antimicrobial activity [2,3]. Fluoroquinolones are of concern because they are widely used and are not readily biodegradable by microorganisms [4]. Fluoroquinolones have been detected in stream and river water [58]. The seven FQs in this studyCIP, LOM, OFL, LEV, CLI, enrooxacin (ENR), and umequine (FLU)represent a range of compounds from different generations and uses. Several, including CIP, LOM, OFL, and LEV, are human-use antibiotics, with CIP and LEV currently being among the most widely used antibiotics in the United States (www.rxlist.com/ top200.htm). Clinaoxacin is an experimental drug and is not currently on the U.S. market. Last, ENR and FLU are used in veterinary medicine, and FLU is used in European aquaculture.
* To whom correspondence may be addressed (mlydy@siu.edu). 423

Limited research has been conducted on the aquatic toxicity of FQ antibiotics. Only a couple of common FQ antibiotics have been investigated, with data available for only a few species. No studies have examined a full complement of FQ antibiotics in terms of toxicity to a variety of nontarget species. Therefore, the objectives of this research include determining the toxicity of seven FQ antibiotics to Pseudokirchneriella subcapitata (formerly Selenastrum capricornutum) and Microcystis aeruginosa, algal species from two divisions (Chlorophyceae and Cyanophyceae), Lemna minor (duckweed), Dapnia magna (waterea), and Pimephales promelas (fathead minnow); and comparing the toxicity results with literaturederived estimated environmental concentrations (EECs) to investigate potential hazards FQ antibiotics may pose to the environment.
MATERIALS AND METHODS

Chemicals
The seven FQ antibiotics used in this study were chosen based, in part, on their availability from the manufacturer and their use in human and veterinary medicine. Information about each compound can be found in Table 1 and chemical structures are illustrated in Figure 1.

Organisms
The toxicity of FQ antibiotics was determined in a variety of aquatic organisms. The species chosen for this study are environmentally important, have widespread distributions, are easy to culture, and have been used in previous studies [9,10]. The cyanobacterium M. aeruginosa and the green alga P. subcapitata were obtained from the UTEX culture collection (University of Texas, Austin, TX, USA). Microcystis aeruginosa was grown in Allens blue-green medium [11] and P.

424

Environ. Toxicol. Chem. 24, 2005

A.A. Robinson et al.

Fig. 1. Chemical structures of the seven uoroquinolone antibiotics used in this study. Levooxacin and ooxacin are stereoisomers with different congurations at the CCH4 location on the molecule.

subcapitata was grown in Guillards freshwater medium [12,13]. Algae were cultured in 25 150-mm glass test tubes with screw-cap lids by using sterile techniques. Cultures were shaken manually to enhance gas exchange and transferred every two weeks to fresh media. Algal cultures were maintained in environmental growth chambers (Precision Scientic, Winchester, VA, USA) with a 12:12 h light:dark cycle at a temperature of 20C under cool-white uorescent lights (100 mol/m2/s) [14]. Lemna minor was obtained from Van Ness water gardens (Upland, CA, USA) and cultured in duckweed nutrient solution [15]. Cultures were maintained in 1-L beakers and kept in environmental growth chambers with a 16:8 h light:dark cycle (100 mol/m2/s) and a temperature of 25C. Fronds were transferred to new nutrient solution every 3 to 4 d. Daphnia magna was obtained from Wright State University (Dayton, OH, USA) and cultured in 1-L beakers containing hard, reconstituted water. Cultures were fed YCT (yeast, Cerophyl, and trout chow mix, Ralston-Purina, St. Louis, MO, USA) and P. subcapitata and kept in environmental chambers (16:8 h light:dark cycle and 25C). Pimephales promelas was obtained from a local aquaculture facility (Logan Hollow, Murphysboro, IL, USA) and cultured according to standard U.S. Environmental Protection Agency methods [16]. Brood stocks of fathead minnows were

held in a recirculating system with carbon-ltered dechlorinated tap water at 25C and a 16:8 h light:dark cycle. Brood stocks were fed three times daily, once with frozen brine shrimp (Artemia salina) and twice with pelleted sh food (Zeigler Bros, Gardners, PA, USA). Ten-gallon aquaria equipped with spawning substrates were used for breeding. The ratio of male to female fathead minnows per tank was 1: 5. Once eggs were laid on the spawning substrate, they were removed and placed in a separate container. Two drops of methylene blue was added as a fungicide. Before hatching, the substrate was moved to fresh water and larvae 24 h old were used in toxicity tests.

Toxicity testing
Algal toxicity tests were performed in 25 150-mm glass test tubes with screw-cap lids. Experiments were set up with 10 replicates with ve different concentrations and a negative control. Growth medium was autoclaved, dosed, and 40 ml was dispensed into test tubes and inoculated with 0.25 ml of P. subcapitata or 1 ml of M. aeruginosa cell suspension. All algal transfer procedures were performed aseptically under a laminar-ow hood. Because of the high chemical concentrations used in the tests with P. subcapitata, neat compound was added directly to the media and diluted to obtain the appropriate concentrations. Tests with M. aeruginosa were

Table 1. A list of uoroquinolone (FQ) antibiotics used in the toxicity tests, where they were obtained, their Chemical Abstracts Service (CAS) numbers, purity from the manufacturer, and dissociation constant (pKa) values. The pKa values were taken from Nowara et al. [33] and Holten Lu tzhft et al. [25] FQ antibiotic Ciprooxacin Levooxacin Ooxacin Lomeoxacin Enrooxacin Clinaoxacin Flumequine Source Serologicals Proteins, Kankakee, IL, USA R.W. Johnson Pharmacy Research Institute, Raritan, NJ, USA Sigma-Aldrich, Oakville, ON, Canada Sigma-Aldrich Bayer, Shawnee Mission, KS, USA Pzer, New York, NY, USA Sigma-Aldrich CAS no. 86393-32-0 100986-85-4 82419-36-1 98079-51-7 93106-60-6 105956-99-8 42835-25-6 Purity (%) 98 100 99 100 98 90 96 pKa 6.4 6.2 6.2 6.0 6.3 6.5 6.4

Toxicity of uoroquinolone antibiotics to aquatic organisms

Environ. Toxicol. Chem. 24, 2005

425

dosed from a stock solution with deionized water as the solvent. The same environmental conditions used for algal culturing were used for the toxicity tests. Test tubes were manually shaken daily throughout each test. In vivo chlorophyll a uorescence was measured on days 0, 1, and 3 for P. subcapitata and days 0, 1, 3, and 5 for M. aeruginosa. Because of slower growth rates in M. aeruginosa, it was necessary to obtain an additional uorometric reading for this species. Fluorescence measurements were taken after vortexing the test tube and directly inserting the tube into a Turner Designs uorometer (model TD-700, Sunnyvale, CA, USA) using a 340- to 500nm excitation lter and a 665-nm emission lter. Growth rates were determined as cell doublings per day and were based on the uorometric readings using the following formula: ln Ft2 ln Ft1/(t2 t1)ln 2 where Ft2 is the uorescence reading at time 2, Ft1 is the uorescence reading at time 1, and t is time [12]. This is a simple and nondestructive method that has been used in several studies as a reliable method for determining growth rates [9,13,14]. Methods for L. minor followed procedures outlined in Wang [15] and Youngman et al. [17]. Two- and three-frond colonies, for a total of 10 to 15 fronds, were placed in 100ml beakers with 50 ml of duckweed nutrient solution. Experiments were set up in triplicate with four concentrations and a negative control and placed in an environmental growth chamber (25C and 16:8 h light:dark cycle) for 7 d. The nutrient solution in each beaker was dosed with a stock solution with deionized water as the solvent. Duckweed tests were performed by using daily static renewal to maintain constant chemical concentrations in the media. At the end of 7 d, fronds were counted. The reproduction rate was then calculated by subtracting the starting number of fronds from the nal number, then dividing by the initial number, resulting in number of new fronds/initial frond/7 d. The 48-h static tests with D. magna were performed according to U.S. Environmental Protection Agency methods [18]. Hard, reconstituted water was used for this test and the test consisted of three replicates at one concentration (10 mg/ L) and a negative control. Analytical-grade (neat) compound was added directly to the water. The concentration of 10 mg/ L was chosen because it is approximately two orders of magnitude higher than the highest reported concentration in hospital wastewaters (125 g/L) [19]. If no response occurred in the D. magna at this concentration, then no further testing was deemed necessary. Ten neonates (24 h old) were placed in 50 ml of water in 250-ml beakers. At the end of 48 h, the number of neonates that died was determined by lack of movement of the body or appendages when gently prodded [18]. Fathead minnow larval survival and growth toxicity tests were performed by using standard U.S. Environmental Protection Agency methods [20]. Ten larval sh (24h old) were placed into 250 ml of moderately hard, reconstituted water in 600-ml beakers. Tests were performed with four replicates at one concentration (10 mg/L) and a negative control. Additional testing with CLI at lower concentrations (0.25, 0.5, 1, and 2 mg/L) was needed to determine the lowest no-observed-effect concentration (NOEC). Beakers were placed into environmental chambers at 25C with a photoperiod of 16:8 h light: dark. At 24-h time intervals, the number of dead sh was determined, the water was renewed, and dead sh were removed from the beakers. During the static renewal process,

water in the beakers was removed until 7 to 10 mm remained and dosed water was added slowly by pouring down the inside of the beaker to avoid stress and injury to the larvae. This resulted in approximately 80% of the water being renewed daily. Larvae were fed 0.1 ml of concentrated Artemia nauplii twice a day, except on the last day of the test. On day 7, surviving sh from each replicate were euthanized, dried, and weighed collectively. The weight was divided by the total number of sh in the replicate to obtain a mean dry weight for that replicate.

Analytical methods
During each toxicity test, FQ concentrations were measured to ensure accurate initial dosing and to determine the stability of the compounds during the toxicity test. During static testing (algae and daphnia), the initial concentration for each test solution was conrmed and three test chambers for each concentration were randomly chosen after the test and analyzed. During the toxicity tests with daily static renewal (fathead minnows and duckweed), concentrations were checked for all test chambers before and after static renewal at least once during the experiment. Aqueous samples were directly analyzed by using liquid chromatography with uorescence and ultraviolet detection based on modication of previous methods [6,21]. Analysis was performed by using an Agilent 1100 series liquid chromatograph equipped with a binary pump, Zorbax SB-C18 column (4.6 150 mm, 5-m particles), and two detectors (ultraviolet and programmable uorescence), in series (all liquid chromatography equipment was from Agilent, Palo Alto, CA, USA). The elution process started with 8% acetonitrile and 92% potassium phosphate buffer (5 mM, pH 3). After 2 min, the percent acetonitrile was set on a gradient to reach 75% at 8 min, and was held steady for 3 min. All compounds were analyzed by using ultraviolet detection at 292 nm, which calibrated linearly from 50 to 10,000 g/L. Fluorescence detection utilized the following optimized excitation and emission wavelengths: CIP, ENR, and LOMexcitation 278 nm and emission 445 nm; FLUexcitation 312 nm and emission 366 nm; OFL, LEV, and CLIexcitation 279 nm and emission 500 nm. The linear range with uorescence detection was between 2 and 1,000 g/L, except for CLI and FLU, whose linearity on uorescence detection only extended down to 20 g/L. Before analysis, the pH of all samples was adjusted to 3, followed by ltration with a 25-mm polyacrylate lter (Acro disk, Pall Gilman, New York, NY, USA) to remove algal cells and other particulate matter that could interfere with analysis. The ltration resulted in less than a 3% loss of analytes. Samples were then analyzed within 24 h of the end of the test.

Statistical methods
Algal (growth and reproduction) and duckweed (reproduction) effective concentrations (EC50s) and their 95% condence intervals were determined by a log-logistic model by using nonlinear regression analysis (PRC NLIN) in SAS [22]. The following equation [23] was used in this model:

y [ ( )]/{1 exp[ log(x/)]}

where y is the growth and reproduction rate, is the bottom end of the sigmoidal curve (in this study, the bottom end of the curve is expected to be zero for growth or reproduction rates; therefore, was set to zero), is the growth and reproduction rate of the control group, is the slope, x is the

426

Environ. Toxicol. Chem. 24, 2005

A.A. Robinson et al.

Table 2. Toxicity, statistical parameters, and analytical data for Microcystis aeruginosa, Pseudokirchneriella subcapitata, and Lemna minor. Median effective concentrations (EC50s) are reported in g/L with condence intervals (CI) reported in parentheses. Alpha is the maximum growth rate observed in the test and beta is the slope of the doseresponse curve. The % of expected is the concentrations that were analyzed at the beginning of a test or 24-h static renewal period and the % recovery is the concentrations analyzed at the end of a test or 24-h static renewal period Antibiotic Levooxacin Ciprooxacin Ooxacin Enrooxacin Levooxacin Clinaoxacin Clinaoxacin Lomeoxacin Enrooxacin Ooxacin Lomeoxacin Ciprooxacin Clinaoxacin Flumequine Flumequine Enrooxacin Flumequine Levooxacin Ooxacin Ciprooxacin Lomeoxacin Species EC50 (95% CI) 7.9 17 21 49 51 62 103 106 114 126 186 203 1,100 1,960 2,470 3,100 5,000 7,400 12,100 18,700 22,700 (6.49.4) (1420) (1824) (4156) (8.694) (21103) (86120) (45167) (84143) (52201) (172200) (41364) (9301,300) (1,7602,160) (1,6503,300) (2,6003,600) (4,8005,200) (6,4008,400) (10,40013,700) (16,20021,200) (19,90025,500) Alpha 0.47 0.43 0.54 0.56 4.16 4.19 0.65 7.62 7.66 4.27 0.99 4.59 1.63 0.46 7.27 1.54 1.35 1.77 1.54 1.67 1.41 Beta 1.81 3.42 1.92 2.33 0.77 1.14 3.47 2.86 2.59 1.09 5.12 1.52 2.56 5.55 1.48 1.23 2.50 1.19 1.47 0.96 1.35 % of Expected 8792 9097 108114 95115 98102 86103 106111 9296 8694 9399 100104 8892 8994 109114 92107 95107 98109 8587 103106 106112 88103 % Recovery 1532 10 1445 20 97120 4059 35 2244 7686 95104 11 8586 8 84102 9798 521 2281 6580 8088 2075 5

M. aeruginosa M. aeruginosa M. aeruginosa M. aeruginosa L. minor L. minor M. aeruginosa L. minor L. minor L. minor M. aeruginosa L. minor P. subcapitata M. aeruginosa L. minor P. subcapitata P. subcapitata P. subcapitata P. subcapitata P. subcapitata P. subcapitata

chemical concentration, and is the midpoint of the curve representing the EC50. The NOEC values for daphnia (survival) and fathead minnows (survival and growth) were determined by the test concentration at which there was 10% or less mortality. Ten percent mortality was set as the acceptable level in the controls. Analysis of variance and Dunnets multiple comparison test were used to analyze dry weights of larval sh [22].

were less pronounced with FLU because M. aeruginosa was only 2.5 times more sensitive than P. subcapitata. Chemical concentrations of the FQs analyzed at the beginning of the tests were all within 15% of expected concentrations (Table 2). The percent recovery at the end of the test was low for most compounds, especially in toxicity tests with M. aeruginosa (Table 2). Lemna minor For L. minor, six of the seven FQs had similar EC50 values, ranging from 53 to 203 g/L. Flumequine was less toxic than the other FQs, with an EC50 value of 2,470 g/L (Table 2). The order of most toxic FQ to the least for L. minor was LEV, CLI, LOM, ENR, OFL, CIP, and FLU. Chlorosis (yellowing of fronds) was observed in L. minor exposed to all FQs at the higher test concentrations starting at 125 g/L. The exception was FLU, at which chlorosis was observed at 3 mg/L. Chlorosis was mainly observed to occur in new fronds produced during the course of the experiment. Initial FQ concentrations were all within 15% of expected concentrations. Percent recovery at the end of a 24-h static renewal period was at least 85%, with the exception of CLI, ENR, and LOM, which had lower recoveries (Table 2). Daphnia magna All FQs tested caused less than 10% mortality in D. magna at a concentration of 10 mg/L. However, in FLU and CLI treatments, unusual lethargic behavior was noted. Approximately 10% of the D. magna neonates swam slowly or only moved appendages when gently prodded. All FQ water concentrations analyzed at the beginning and end of the 48-h tests were within 10% of expected concentrations except for CLI. The ending water concentrations for CLI were 77 to 82% of expected. Pimephales promelas Six out of seven FQs exposed to larval fathead minnows caused less than 8% mortality at a water concentration of 10

Hazard quotients
To assess the environmental hazard posed by FQ antibiotics to the aquatic organisms tested in this study, the concentration for each measured endpoint (median lethal concentration, EC50, or NOEC) was divided by the EEC to obtain a hazard quotient (HQ). Two different EECs were calculated, a high EEC value of 100 g/L was chosen to represent the worstcase scenario, which would involve direct hospital wastewater input. A moderate EEC value of 1 g/L was chosen to represent the highest concentration that is likely to occur in surface water. If a HQ is greater than one, there is perceived hazard to that organism. The higher the HQ value, the greater the hazard, and the lower the HQ value, the smaller the hazard is to that organism. A safety factor of 10 was added to this value to compensate for intra- and interspecies variability resulting in a benchmark HQ of 0.1.
RESULTS

Microcystis aeruginosa and P. subcapitata The EC50 values for M. aeruginosa and P. subcapitata ranged from 7.9 to 1,960 g/L and 1,100 to 23,000 g/L, respectively (Table 2). The following represent the order of FQs from the most to least toxic for M. aeruginosa: LEV, CIP, OFL, ENR, CLI, LOM, and FLU. For P. subcapitata, the order of most toxic FQ to the least was CLI, ENR, FLU, LEV, OFL, CIP, and LOM. Microcystis aeruginosa was approximately two orders of magnitude more sensitive to six of the seven FQ antibiotics than P. subcapitata. Differences in sensitivity

Toxicity of uoroquinolone antibiotics to aquatic organisms Table 3. Hazard quotients (HQs) for Microcystis aeruginosa, Pseudokirchneriella subcapitata, Lemna minor, Daphnia magna, and Pimephales promelas when using a high estimated environmental concentration (EEC) (100 g/L) and a moderate EEC (1 g/L). Italicized values represent high perceived risk to the organism (HQ 0.1) Antibiotic Ciprooxacin Levooxacin Ooxacin Lomeoxacin Enrooxacin Clinaoxacin Flumequine Ciprooxacin Levooxacin Ooxacin Lomeoxacin Enrooxacin Clinaoxacin Flumequine Ciprooxacin Levooxacin Ooxacin Lomeoxacin Enrooxacin Clinaoxacin Flumequine All 7 FQsa Ciprooxacin Levooxacin Ooxacin Lomeoxacin Enrooxacin Clinaoxacin Flumequine
a

Environ. Toxicol. Chem. 24, 2005

427

Species

HQ HQhigh moderate EEC EEC 0.059 0.127 0.048 0.005 0.020 0.010 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.005 0.020 0.008 0.009 0.009 0.016 0.000 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001

M. aeruginosa M. aeruginosa M. aeruginosa M. aeruginosa M. aeruginosa M. aeruginosa M. aeruginosa P. subcapitata P. subcapitata P. subcapitata P. subcapitata P. subcapitata P. subcapitata P. subcapitata L. minor L. minor L. minor L. minor L. minor L. minor L. minor D. magna P. promelas P. promelas P. promelas P. promelas P. promelas P. promelas P. promelas

5.88 12.7 4.76 0.538 2.04 0.971 0.051 0.005 0.014 0.008 0.004 0.032 0.091 0.020 0.493 1.96 0.794 0.943 0.877 1.61 0.040 0.01 0.01 0.01 0.01 0.01 0.01 0.05 0.01

FQ uoroquinolone.

mg/L. However, fathead minnows treated with CLI at 10 mg/ L displayed 100% mortality. Subsequent testing with CLI revealed no mortality at 0.25, 0.5, 1, or 2 mg/L. No signicant differences were found in dry weight of the sh in the LOM, ENR, CLI, and FLU treatments, with mean masses 95 to 119% of control sh. However, sh treated with CIP, LEV, and OFL weighed signicantly more than the control (p 0.05), with masses of 122, 138, and 121% of the control mass, respectively. Fluoroquinolone water concentrations analyzed at the beginning and end of a 24-h static renewal period ranged from 88 to 128% of expected concentration for all tests, with the exception of the ending concentration of LOM, which dropped to 34 to 74% of expected.

Hazard quotients
For the moderate EEC, only one HQ exceeded the benchmark value of 0.1 (Table 3). The lone exceeding value was for LEV toxicity to M. aeruginosa. Several other moderate HQs were close to the benchmark for this species. In contrast, every HQ for FQ toxicity to M. aeruginosa and L. minor exceeded the high EEC benchmark except for FLU. The HQs for P. subcapitata, D. magna, and P. promelas were lower than the benchmark value for each FQ even at the high EEC.
DISCUSSION

cluding P. subcapitata (median EC50 7,400 g/L) and D. magna and P. promelas (NOECs at or near 10,000 g/L). Microcystis aeruginosa was the most sensitive to FQ exposure, with a median EC50 value of 49 g/L. Lemna minor was the next most sensitive organism (median EC50 106 g/L). This nding is similar to that of another multiorganism study that tested CIP on M. aeruginosa, P. subcapitata, D. magna, and Brachydanio rerio (zebrash). Microcystis aeruginosa was found to be the most sensitive and P. subcapitata the next most sensitive organism, whereas D. magna and B. rerio had no acute toxicity at 60 and 100 mg/L, respectively [24]. The vast difference in sensitivity between the two algal species, M. aeruginosa and P. subcapitata, to FQs is consistent with other studies that found M. aeruginosa to be two to three orders of magnitude more sensitive than P. subcapitata to antibiotics [10,25]. Cyanobacteria are prokaryotic, making them structurally similar to bacteria; therefore, M. aeruginosa may be the most susceptible of the tested organisms to the antibiotics mode of action. Interestingly, the FQ antibiotics tested on L. minor in the present study exhibited toxicity nearing that observed for M. aeruginosa. Previous researchers have suggested the possible explanation that chloroplasts are prokaryotic in nature and are thought to have similar metabolic pathways as that of cyanobacteria [26,27]. Therefore, the chloroplasts within L. minor are likely to be susceptible to the toxic mode of action of FQ antibiotics. However, it would follow that similar observations would occur in P. subcapitata, but no such observations of chlorosis were made in this study. In terms of EC50 values, it is unclear why P. subcapitata is less sensitive to FQ activity than L. minor. Although sh were general not affected by FQ exposure even at high levels (2,00010,000 g/L), sh dry weights were signicantly greater as compared to the control for CIP, LEV, and OFL. A possible explanation for this response is that at a concentration of 10 mg/L, these compounds act as growth promoters, reducing the amount of deleterious bacteria in the sh. This response has been noted in livestock operations where antibiotics are used as growth promoters (Animal and Plant Health Inspection Service, http://www.aphis.usda.gov/ vs/ceah/cei/antiresist.antibiouse.pdf). A similar response also was seen in rainbow trout treated with the antibiotic avoparcin [28]. The only FQ antibiotic found to have any deleterious effects on P. promelas was CLI, which produced nearly 100% mortality at 10 mg/L. Subsequent testing with CLI revealed no mortality in P. promelas at a concentration of 2 mg/L. It is uncertain why CLI is the only FQ to produce an adverse effect on P. promelas survival at 10 mg/L. Clinaoxacin also was the most toxic FQ to P. subcapitata and there was some mobility impairment in D. magna. It is interesting to note that CLI is a fourth-generation FQ and has wider microbial activity compared to the other FQs in this study [2]. Correspondingly, FLU, an early-generation FQ with limited antimicrobial activity, was the least toxic to L. minor and M. aeruginosa.

Comparison to previous studies

Previous algal EC50 values have been reported for FLU and CIP. For FLU, P. subcapitata and M. aeruginosa have reported EC50 values of 5 and 0.159 mg/L, respectively [25]. For CIP, P. subcapitata and M. aeruginosa have reported EC50 values of 2.97 and 0.005 mg/L, respectively [24]. These

Sensitivity of species
The FQ antibiotics investigated in this study demonstrated very limited toxicity to three of the ve species tested, in-

428

Environ. Toxicol. Chem. 24, 2005

A.A. Robinson et al.

results are similar to those reported in this study for FLU with P. subcapitata (5 mg/L) and for CIP with M. aeruginosa (0.017 mg/L). However, the EC50 values for FLU with M. aeruginosa in the previously reported study (0.159 mg/L) contrasts with that of the current study (1.96 mg/L). This may be due to differences in test methods. The previous study [25] performed M. aeruginosa tests for 7 d as opposed to 5 d in the current study. The previously reported EC50 value for P. subcapitata and CIP (2.97 mg/L) also is not consistent with that of the current study (19 mg/L), although both values are quite high as compared to likely environmental concentrations. The EC50 values for L. minor in this study are different from a previous study, which reported EC50 values for Lemna gibba of 116, 194, 697, and 653 g/L for LOM, LEV, CIP, and OFL, respectively [27]. These values, except for LOM, are greater than what was reported in this study, which were 106, 51, 203, and 126 g/L, respectively. This is likely due to differences between duckweed species. Nevertheless, the difference in EC50 values is not great because both sets of values are within a factor of ve. The chlorosis observed in new fronds at the higher concentrations in this study is consistent with previously reported observations [27]. Species of Lemna appear to be more sensitive than other aquatic plants. The aquatic plant Lythrum salicaria L. exhibited hormesis (increased growth compared to controls) at FLU concentrations of 50 to 5,000 g/L and toxicity was observed at 100 mg/L [29]. In the present study, D. magna was not sensitive to FQ exposure. This is consistent with the NOEC value of 60 mg/ L reported for D. magna with CIP [24]. In another study, a rst-generation quinolone, oxalinic acid, was tested on D. magna neonates and a 48-h EC50 value of 4.6 mg/L was calculated, indicating that some FQs may be toxic to Daphnia at extremely high concentrations [30]. Previous research with a different crustacean, nauplii of Artemia salina, found a 48h EC50 value of 308 mg/L [31]. Cyst hatchability tests also were performed with 0.1 to 1 mg/L of FLU and these tests found no effect on hatching rate of A. salina; however, 74% of the hatched nauplii lacked orange pigmentation and were transparent [31]. No studies were found to compare FQ toxicity with P. promelas. However, a NOEC for B. rerio (zebrash) with CIP was reported to be 100 mg/L [24]. Fluoroquinolone antibiotics such as FLU have limited toxicity to adult sh, as demonstrated by their use in aquaculture.

tions. Examination of unpublished data from our laboratory suggests that this may be due to FQs adsorbing to algal cells. Lower FQ concentrations caused less inhibition in the algae, resulting in more cells with which the antibiotics could adsorb. This is also supported by reports that FQs bind to particulate matter [33]. Because initial water concentrations of the FQs were used to obtain EC50 values in this study, these values likely underestimate toxicity because they assume constant concentrations throughout the length of the test. However, it is difcult to determine to what extent these values underestimate toxicity. A recent study addressed this issue by analyzing copper concentrations throughout a 72-h algal toxicity test and developing a model to demonstrate the amount that toxicity values may be underestimated according to different concentration-decline scenarios [34]. That study found that copper toxicity could be underestimated by as much as a factor of 2 and modeling exercises indicate that rapidly declining concentrations may underestimate toxicity by as much as a factor of 50. In the current study, all the FQs except FLU had greater concentration drops in tests with M. aeruginosa than in tests with P. subcapitata, but this is likely due to the fact that M. aeruginosa tests were 2 d longer in duration. Therefore, toxicity values for M. aeruginosa may be underestimated more than toxicity values for P. subcapitata.

Environmental concentrations of FQ antibiotics

Environmental concentrations of FQs are reported to be in the low g/L to ng/L range. Reports have come from three different types of sampling sites, including hospital wastewater, municipal wastewater, and receiving river water. Theoretical concentrations of CIP in these three sampling sites were estimated to be 2 to 30, 0.6, and 0.06 g/L, respectively [4]. Measured environmental concentrations were similar to these theoretical values. In German hospital wastewater, CIP was detected in all 24 samples at concentrations ranging from 0.7 to 124.5 g/L, with a median of 3.1 g/L [19]. In an earlier study, Hartmann et al. [35] reported CIP in hospital wastewater at concentrations ranging from 3 to 87 g/L. The high EEC value for this study was set at 100 g/L to conservatively represent the worst-case scenario of direct exposure to hospital wastewater with little dilution. In comparison to hospital wastewater, municipal wastewater efuent and receiving river water have lower FQ concentrations. In municipal wastewater efuent, CIP was detected at concentrations ranging from 0.062 to 0.106 g/L, with a mean value of 0.072 g/L [36]. The FQ concentrations in river water were 0.005 to 0.018 g/L in the Glatt River, Switzerland [36], and 0.014 to 0.026 g/L in the Po and Lambro Rivers, Italy [8]. In streams across the United States, two FQ antibiotics were detected, CIP and noroxacin, with maximum concentrations of 0.03 and 0.12 g/L, respectively [7]. The moderate EEC value for this study was set at 1.0 g/L to conservatively represent these surface water concentrations. Because of the limited number of sites and analyte lists, this EEC value should be considered a rough estimate. However, it is likely a conservative estimate, based on our current knowledge.

Stability of FQ antibiotics in test systems

Depending on conditions, FQ antibiotics can have short half-lives in aqueous systems, primarily because of photodegradation [32] and sorption to particulate matter [33]. To ensure constant FQ exposure, testing was performed by using static renewal whenever possible. However, because of the inherent difculties in conducting static renewal tests with algae, static tests were performed with chemical analysis at the beginning and end of the test to get a measure of antibiotic stability within the test. The alternative method of static renewal tests would require ltering and centrifuging algal cells, which could lead to cell damage and lower growth rate in the controls. For many of the algal toxicity tests, examination of analytical data shows a large decline in FQ concentrations at the end of the tests. The large range of FQ recoveries reported in Table 2 was dependent on initial concentrations. For example, lower concentrations fell more rapidly than did higher concentra-

Hazard assessment
Hazard quotients were calculated to help dene the problem statement in future risk assessment efforts by determining which environmental components may potentially be exposed

Toxicity of uoroquinolone antibiotics to aquatic organisms

Environ. Toxicol. Chem. 24, 2005

429

to toxic concentrations of FQ antibiotics. From this study, it is evident that P. subcapitata, D. magna, and P. promelas are not likely to be at risk because of environmental FQ exposure. Lemna minor would only be at risk in worst-case scenarios, whereas M. aeruginosa may be the most likely to be at risk to FQ exposure, especially if environmental concentrations are persistent. It is possible for FQs to have a pseudo-persistence quality [35] due to continual input of sewage treatment into the environment. This phenomenon may negate other processes that could eliminate FQs from the system, such as photodegradation [32] and sorption to particulate matter [33] and sediment [37]. Because FQ concentrations in toxicity tests with M. aeruginosa dropped signicantly by the end of the 5-d tests, these toxicity data may underestimate the risk these compounds could pose in the environment. This study assessed the toxic effects and environmental risk of FQ antibiotics to aquatic organisms. By using a multiorganismal approach, we were able to elucidate some of the species most at risk to FQs in the environment. However, because of the possibility of FQs binding to particulate matter [33] and sediment [37], it is of interest for future research to study the fate and effects of these chemicals in benthic environments. Because the greatest toxicity observed in this study was with prokaryotic organisms, future research also should look into the effects of FQs on aquatic bacterial communities, such as microbial resistance and the possible disruption of the aquatic nutrient cycle. Finally, because a variety of pharmaceuticals may enter surface water, the effects of these mixtures to aquatic organisms may be an important aspect to consider for future research.
AcknowledgementThis research was funded by a U.S. Environmental Protection Agency Science to Achieve Results (STAR) grant (R-82900801-0). We wish to thank Bayer Corporation, R.W. Johnson, and Pzer for providing some of the antibiotics.
REFERENCES 1. Hooper DC. 1995. Quinolone mode of action. Drugs 49(Suppl 2):1015. 2. Appelbaum PC. 1995. Quinolone activity against anaerobes: Microbiological aspects. Drugs 49(Suppl 2):7680. 3. King DE, Malone R, Lilley SH. 2000. New classication and update on the quinolone antibiotics. Am Fam Physician 61:2741 2748. 4. Al-Ahmad A, Daschner FD, Ku mmerer K. 1999. Biodegradability of cefotiam, ciprooxacin, meropenem, penicillin G, and sulfamethoxazole and inhibition of waste water bacteria. Arch Environ Contam Toxicol 37:158163. 5. Campagnolo ER, Johnson KR, Karpati A, Rubin CS, Kolpin DW, Meyer MT, Esteban E, Currier RW, Smith K, Thu KM, McGeehin M. 2002. Antimicrobial residues in animal waste and water resources proximal to large-scale swine and poultry feeding operations. Sci Total Environ 299:8995. 6. Golet EM, Alder AC, Giger W. 2002. Environmental exposure and risk assessment of uoroquinolone antibacterial agents in wastewater and river water of the Glatt Valley watershed, Switzerland. Environ Sci Technol 36:36453651. 7. Kolpin DW, Furlong ET, Meyer MT, Thurman EM, Zaugg SD, Barber LB, Buxton HT. 2002. Pharmaceuticals, hormones, and other organic wastewater contaminants in U.S. streams, 1999 2000: A national reconnaissance. Environ Sci Technol 36:1202 1211. 8. Calamari D, Zuccato E, Castiglioni S, Bagnati R, Fanelli R. 2003. Strategic survey of therapeutic drugs in the rivers Po and Lambro in northern Italy. Environ Sci Technol 37:12411248. 9. Fairchild JF, Ruessler DS, Carlson AR. 1998. Comparative sensitivity of ve species of macrophytes and six species of algae to atrazine, metribuzin, alachlor, and metolachlor. Environ Toxicol Chem 17:18301834.

10. Halling-Srensen B. 2000. Algal toxicity of antibacterial agents used in intensive farming. Chemosphere 40:731739. 11. Fogg GE, Stewart WDP, Fay P, Walsby AE. 1973. The BlueGreen Algae. Academic, New York, NY, USA. 12. Guillard RRL. 1975. Culture of phytoplankton for feeding marine invertebrates. In Smith WL, Chanley MH, eds, Culture of Marine Invertebrate Animals. Plenum, New York, NY, USA, pp 2960. 13. Nelson KJ, Hoagland KD, Siegfried BD. 1999. Chronic effects of atrazine on tolerance of a benthic diatom. Environ Toxicol Chem 18:10381045. 14. Taylor J. 2001. Effects of binary combinations of herbicides on freshwater algae. MS thesis. University of NebraskaLincoln, Lincoln, NE, USA. 15. Wang W. 1990. Literature review on duckweed toxicity testing. Environ Res 52:722. 16. U.S. Environmental Protection Agency. 1993. Methods for measuring the acute toxicity of efuents and receiving waters to freshwater and marine organisms, 4th ed. EPA 600/4-90/027F. Ofce of Research and Development, Cincinnati, OH. 17. Youngman AL, Williams TL, Lydy MJ. 1998. Utilization of a duckweed bioassay to evaluate leaching of heavy metals in smelter contaminated soils. In Little EE, DeLonay AJ, Greenberg BM, eds, Environmental Toxicology and Risk Assessment, Vol 7. STP 1333. American Society for Testing and Materials, Philadelphia, PA, pp 227235. 18. U.S. Environmental Protection Agency. 1988. Protocols for short term toxicity screening of hazardous waste sites. EPA 600/3-88/ 029. Corvallis, OR. 19. Hartmann A, Golet EM, Gartiser S, Alder AC, Koller T, Widmer RM. 1999. Primary DNA damage but not mutagenicity correlates with ciprooxacin concentrations in German hospital wastewaters. Arch Environ Contam Toxicol 36:115119. 20. U.S. Environmental Protection Agency. 1994. Short-term methods for estimating the chronic toxicity of efuents and receiving waters to freshwater organisms, 3rd ed. EPA/600/4-91/002. Ofce of Research and Development, Cincinnati, OH. 21. Yorke JC, Froc P. 2000. Quantitation of nine quinolones in chicken tissues by high-performance liquid chromatography with uorescence detection. J Chromatogr A 882:6377. 22. SAS Institute. 2000. SAS/STAT Users Guide, 8th ed. Cary, NC, USA. 23. Schabenberger O, Tharp BE, Kells JJ, Penner D. 1999. Statistical tests for hormesis and effective dosages in herbicide dose response. Agron J 91:713721. 24. Halling-Srensen B, Holten Lu tzhft H-C, Andersen HR, Ingerslev F. 2000. Environmental risk assessment of antibiotics: Comparison of mecillinam, trimethoprim, and ciprooxacin. J Antimicrob Chemother 46(S1):5358. 25. Holten Lu tzhft HC, Halling-Srensen B, Jrgensen SE. 1999. Algal toxicity of antibacterial agents applied in Danish sh farming. Arch Environ Contam Toxicol 36:16. 26. McFadden GI, Roos DS. 1999. Apicomplexan plastids as drug targets. Trends Microbiol 7:328333. 27. Brain RA, Johnson DJ, Richards SM, Sanderson H, Sibley PK, Solomon KR. 2004. Effects of 25 pharmaceutical compounds to Lemna minor using a seven-day static renewal test. Environ Toxicol Chem 23:371382. 28. Zoccarato I, Gasco L, Calvi SL, Fortina R, Bianchini ML, Rollin X. 1995. Effect of dietary avoparcin on performances and carcass composition in rainbow trout, Oncorhynchus mykiss (Walbaum). Aquacult Res 26:361366. 29. Migliore L, Cozzolino S, Fiori M. 2000. Phytotoxicity to and uptake of umequine used in intensive aquaculture on the aquatic weed, Lythrum salicaria L. Chemosphere 40:741750. 30. Wollenberger L, Halling-Srensen B, Kusk KO. 2000. Acute and chronic toxicity of veterinary antibiotics to Daphnia magna. Chemosphere 40:723730. 31. Migliore L, Civitareale C, Brambilla G, Dojmi di Delupis G. 1997. Toxicity of several important agricultural antibiotics to Artemia. Water Res 31:18011806. 32. Burhenne J, Ludwig M, Nikoloudis P, Spiteller M. 1997. Primary photoproducts and half-lives. Environ Sci Pollut Res 4:1015. 33. Nowara A, Burhenne J, Spiteller M. 1997. Binding of uoroquinolone carboxylic acid derivatives to clay minerals. J Agric Food Chem 45:14591463.

430

Environ. Toxicol. Chem. 24, 2005

A.A. Robinson et al. 36. Golet EM, Alder AC, Hartmann A, Ternes TA, Giger W. 2001. Trace determination of uoroquinolone antibacterial agents in urban wastewater by solid-phase extraction and liquid chromatography with uorescence detection. Anal Chem 73:36323638. 37. Hektoen H, Berge JA, Hormazabal V, Yndestad M. 1995. Persistence of antibacterial agents in marine sediments. Aquaculture 133:175184.

34. Simpson SL, Roland MGE, Stauber JL, Batley GE. 2003. Effect of declining toxicant concentrations on algal bioassay endpoints. Environ Toxicol Chem 22:20732079. 35. Hartmann A, Alder AC, Koller T, Widmer RM. 1998. Identication of uoroquinolone antibiotics as the main source of umuC genotoxicity in native hospital wastewater. Environ Toxicol Chem 17:377382.