Stock solution preparation
ATV (in the form of calcium trihydrate, more than 99.9 %, European Pharmacopeia Reference Standard, purchased from Merck, Germany) solution at a concentration of 50 mg/l was prepared by dissolution in double-deionised water (Smart2Pure6UV/UF, Thermo Scientific). FeCl3 (p.a., Lachema, Czech Republic) was added to the ATV solution before irradiation to make a final concentration of 5 mg Fe/l.
Irradiation
Irradiation was carried out in a Rayonet R100 reactor equipped with twelve RPR 3000Å lamps emitting light at a wavelength range of 250 – 350 nm; to exclude wavelengths below 300 nm samples were irradiated in optical glass cuvettes covered with PTFE lids.
Actinometric measurement using ferrioxalate actinometry resulted in the value of 8.5×10-8 ein s-1 per the cuvette surface area.
For further experiments with irradiated samples, the P-ATV was used.
Analyses of samples
HPLC analysis
The extent of ATV photodegradation was determined by HPLC (ThermoScientific Dionex Ultimate system 3000 (USA), column Phenomenex Kinetex® 5 µm EVO C18, 30 × 2.1 mm, mobile phase water with 0.001 % HCOOH and acetonitrile 0.6:0.4 (v/v) with PDA 3000RS spectrophotometric and FLD 3000RS fluorescence detectors.
TOC analysis
Total organic carbon content of both non-irradiated and irradiated ATV samples was determined as nonpurgable total organic carbon after catalytic combustion at 720°C by Shimadzu TOC-L.
HPLC-HRMS/MS analysis
The photoproducts mixture (P-ATV) was analysed using a Thermo Scientific Dionex UltiMate 3000 ultra-high-performance liquid chromatography plus (UHPLC +) instrument equipped with a diode array detector connected to a Bruker Impact HD (Bruker, Billerica, MA, USA) high-resolution mass spectrometer with electrospray ionisation and QTOF analyzers. Separation of photoproducts was performed on Arion Plus C18 column (3 µm; 150 mm × 4.6 mm) using (A) acetonitrile containing 0.1 % formic acid and (B) water + 0.1 % formic acid as a mobile phase with a 0.6 ml/min flow rate. The gradient was as follows: A/B (v/v) 15/85 (0 min), 15/85 (3 min), 100/0 (23 min), 100/0 (28 min), and 15/85 (33 min). The volume of sample injection was 5 µl. The parameters of the ion source and MS/MS experiment are given in Table 1. The obtained spectra were calibrated using the Lock Mass Tuning Mix ES-TOF internal calibration standard (622 Da) and sodium formate clusters at the beginning of each analysis using Bruker Compass DataAnalysis 5.1 and Bruker Compass LibraryEditor 5.1 software.
Toxicity assays
The ecotoxicity of both the ATV and P-ATV was evaluated using series of toxicity assays on three model organisms representing different levels of the aquatic food web. Combining the assays on primary producer Desmodesmus subspicatus, cladoceran grazer Daphnia magna and fish Danio rerio provided a more complex picture of potential negative effects of studied substances.
To distinguish the effect of photoproducts from the effect of ATV itself, the amount of the irradiated P-ATV (with the remaining 15 % of the original ATV amount) added to the tested organisms was 6.7 times of the volume of non-irradiated ATV so that the concentrations of ATV in both assay growing media were of the same value.
Desmodesmus toxicity assay
The growth inhibition test based on OECD methodology (OECD 201) was used to assess the toxicity of ATV and P-ATV on algae. The stock solutions of ATV and P-ATV were pre-diluted in Z medium (growth medium used for cyanobacteria and algae cultivation according to Staub 1961) to obtain concentrations of 10 mg/l, resp. 5 mg/l of remaining atorvastatin in irradiated solution. The initial cell concentration of the model organism Desmodesmus subspicatus (obtained from CCALA Třeboň, Czech Republic, https://ccala.butbn.cas.cz/en; strain number BRINKMANN 1953/SAG 86.81) was 4 103 cell/ml. The growth inhibition test was done in 96-well microplates. Six test concentrations ranging from 2.4 10-3 mg/l to 10 mg/l were used for ATV and another six concentrations ranging from 3.05 10-4 mg/l to 1.25 mg/l for P-ATV. Test concentrations were arranged in a geometric series with a factor of 4. As a control, only the Z medium with the same amount of Desmodesmus cells was used. All microplates were incubated in Climacell EVO line incubator tempered to a constant temperature of 23.0 ± 0.1 °C with constant irradiation. The growth of the alga was quantified as a fluorescence signal using a BioTek Cytation 5 microtitration plate reader (excitation wavelength: 485 nm, emission wavelength: 680 nm) every 24 hours for 72 hours. Shaking of the microplate was applied prior to the fluorescence reading.
Daphnia toxicity assays
To evaluate the possible effect of ATV and P-ATV on Daphnia magna, both the short-term acute and long-term chronic toxicity assays were performed (cf. OECD 202 and OECD 211). The stock solutions of ATV and P-ATV were diluted in ADaM (Aachener Daphnien Medium, which is based on synthetic sea salt and analytical grade chemicals added to deionized water according to Klüttgen et al. 1994) to the required concentrations of 0.5 –40 mg/l of ATV and 0.01 – 5 mg/l of remaining ATV in P-ATV mixture. In the acute toxicity assay, two neonates of D. magna not older than 24 hours were introduced into 10 ml of test concentrations and into the ADaM medium serving as a control. Acute toxicity was performed in five replicates of nine concentrations for ATV and ten replicates of seven concentrations for P-ATV, respectively. The experiments were kept in an incubator (temperature 20.2 ± 0.2 °C, photoperiod 16 hours light and 8 hours dark). The daphnids were not fed during the assay and their immobilisation was checked after 24 and 48 hours.
Chronic exposure of ATV and P-ATV was performed according to OECD 211. Female neonates – daphnids younger than 24 hours – were individually introduced into 50 ml of media in three treatments: ATV, P-ATV, and the control. The concentrations used were as follows: the ATV treatment concentration was 200 μg/l and 50 μg/l, and the photodegradation product treatment (P-ATV) was added into the medium in a volume that guaranteed the ATV concentration of 200 μg/l and 50 μg/l, so that any difference in toxicity could be attributed to the photoproducts present in the mixture. All dilutions were done in ADaM and every treatment ran in 15 replicates. The test lasted 21 days and took place in the same incubator with the same conditions as the acute toxicity test above. The daphnids were fed every second day with the green alga Desmodesmus subspicatus (3 mg C/l), the medium was changed every four days. The survival and number of offspring were monitored every other day.
Danio toxicity assay
The toxicity assay on Danio rerio followed the OECD 236 method (2013). The stock solutions of ATV and P-ATV were diluted in ISO water (ISO, 2007) to obtain nine concentrations of 0.0001–25 mg/l of ATV and seven concentrations (0.001–5 mg/l) of remaining ATV in the P-ATV mixture (since the higher concentrations of remaining ATV in the P‑ATV mixture could not be obtained). Newly fertilised eggs were selected under a stereomicroscope and transferred into 2 ml of a particular concentration of ATV or P-ATV in the 24-well microplates. Every microplate represented one concentration resulting in 24 replicates. Pure ISO water was used as a control. The test was performed in an incubator tempered at 24.9 0.2 °C with a photoperiod of 13 hours of light and 11 hours of dark for 96 hours. Solutions of experimental concentrations were exchanged once after 48 hours. Coagulation of fertilised eggs, lack of somite formation, lack of detachment of the tail from the yolk sac, and a slow heartbeat, if present, were recorded as indicators of lethality every 24 hours; morphological and behavioural abnormalities were checked using a stereomicroscope.
Statistical analysis
The growth inhibition of algae, immobilisation of Daphnia, and mortality rate of Danio embryos were calculated as a measure of the effect of both ATV and P-ATV. The IC50 and LC50 values of all three tested organisms were obtained after fitting the non-linear regression using the GraphPad Prism 7.05 software package (GraphPad Software, www.graphpad.com). Differences in the number of juveniles and clutches, interclutch period, and body size in the long-term toxicity assay on D. magna were assessed by non-parametric Kruskal-Wallis test followed by post-hoc comparisons using Statistica 14 (TIBCO Software Inc., 2020). In zebrafish embryos, mortality at 96 hpf, hatching rate and abnormalities for the whole experimental period were assessed by generalized linear model (GLM) with an assumed binomial distribution using R (R Core Team 2023). GLM models including substance (ATV vs. P-ATV), concentration levels and interactions between substance and concentrations were built. The best-fitting model (including substance, concentration and interaction) based on Akaike information criterion (AIC) is presented. The treatment effect on zebrafish mortality, hatching rate and abnormalities was tested by chi square test.