Ready biodegradability of trifluoromethylated phenothiazine drugs ...

2 downloads 0 Views 4MB Size Report
Jun 8, 2012 - RESEARCH ARTICLE. Ready biodegradability of trifluoromethylated phenothiazine drugs, structural elucidation of their aquatic transformation.
Environ Sci Pollut Res (2012) 19:3162–3177 DOI 10.1007/s11356-012-1002-1

RESEARCH ARTICLE

Ready biodegradability of trifluoromethylated phenothiazine drugs, structural elucidation of their aquatic transformation products, and identification of environmental risks studied by LC-MSn and QSAR Christoph Trautwein & Klaus Kümmerer

Received: 15 March 2012 / Accepted: 21 May 2012 / Published online: 8 June 2012 # Springer-Verlag 2012

Abstract The environmental fate of transformation products from organic pollutants such as drugs has become a new research area of increasing interest over the last few years. Whereas in the past mainly parent compounds or their major human metabolites were studied, new questions have arisen what compounds could be formed during incomplete degradation in the aquatic environment and what effects the resulting transformation products might have on nature and mankind. Psychiatric drugs are among the most important prescription drugs worldwide, but so far only little data is provided upon their degradation behavior. This especially accounts for tricyclic antipsychotic drugs of the phenothiazine class. Therefore, the degradation of such drugs was investigated in this study. In this study the aerobic Closed Bottle test (The Organisation for Responsible editor: Philippe Garrigues Electronic supplementary material The online version of this article (doi:10.1007/s11356-012-1002-1) contains supplementary material, which is available to authorized users. C. Trautwein Department of Environmental Health Sciences, University Medical Centre Freiburg, Breisacher Str. 115B, 79106 Freiburg, Germany e-mail: [email protected] C. Trautwein Institute for Biology II, Microbiology, University Freiburg, Schänzlestrasse 1, 79104 Freiburg, Germany C. Trautwein : K. Kümmerer (*) Institute of Sustainable and Environmental Chemistry, Leuphana University Lüneburg, Scharnhorststrasse 1, 21335 Lüneburg, Germany e-mail: [email protected]

Economic Co-operation and Development (OECD) 301D) was used to assess the ready biodegradability of three trifluoromethylated phenothiazine drugs: fluphenazine, triflupromazine, and trifluoperazine. As it is known from literature that phenothiazine drugs can easily form various photolytic transformation products under light exposure, photochemical transformation was also investigated. Since transformation products are usually not available commercially, the calculation of environmental parameters with the aid of quantitative structure activity relationship (QSAR) software was used for first evaluation of these compounds. According to the OECD test guideline, all trifluoromethylated phenothiazines had to be classified as not readily biodegradable. Chromatographic data revealed the formation of some transformation products. Comparing retention time and mass spectrometric data with the analytical results of the light exposure experiments, we found peaks with the same retention time and mass spectra. So these transformation products were not of bacterial, but photolytic, origin and are formed very quickly even under low light doses. A special chromatographic column and solvent gradient along with multiple stage mass spectrometric fragmentation experiments uncovered the presence of, in total, nine photolytic transformation products and allowed for their structural elucidation. Typical modifications of the molecules were sulfoxidation, exocyclic N-oxidation, and transformation of the trifluoromethyl to a carboxylic moiety. The obtained results of the QSAR calculations show that all transformation products are highly mobile in the aquatic environment and elimination through biotic or abiotic pathways cannot be expected. Transformation products of trifluoromethylated phenothiazine drugs have to be expected in the aquatic environment, yet nothing is known about their toxicological properties. Therefore, further risk assessment upon these drugs and their fate is strongly recommended.

Environ Sci Pollut Res (2012) 19:3162–3177

Keywords Biodegradation . Persistence . Photolysis . Toxicity . Fluphenazine . Triflupromazine . Trifluoperazine

Introduction History and pharmacology The year 1952 was a hallmark in the history of pharmacology: the clinical introduction of the first modern antipsychotic drug chlorpromazine helped for the first time to combat successfully neuroleptic diseases like schizophrenia and mania (López-Muñoz et al. 2005). Chlorpromazine replaced heavy impact therapies like electro and insulin shocks or lobotomy surgery causing permanent brain damage (Buschmann et al. 2007). Chlorpromazine and related phenothiazine pharmaceuticals are, even after more than 50 years of usage, still common in many countries. The neuroleptic potency of phenothiazines depends strongly on their aromatic ring substitutes. Unsubstituted aliphatic phenothiazines like promazine have a low potency. By contrast, piperidine and piperazine side chains at position 10 increase the antipsychotic effectiveness as well as substitution with Cl and CF3 at position 2 of the aromatic heterocycle (Meyer and Quenzer 2004). Trifluoromethylated phenothiazine pharmaceuticals are classical antipsychotic drugs, which are mainly used in the treatment of schizophrenia and other mental disorders. The parent compounds and some of their metabolites interfere with the dopamine system and block dopamine D1 and D2 receptors. Blockade of D2 receptors in the striatum is thought to be responsible for extrapyramidal motor effects, which have been reported frequently (Harasko-van der Meer et al. 1993). Besides extrapyramidal motor disturbances, photosensitivity was reported for trifluoromethylated phenothiazine pharmaceuticals (Eberlein-Koenig et al. 1997). Fluphenazine (Prolixin®, Fig. 3—first line) was clinically introduced in the year 1959 and used since then against schizophrenia, mania, and other mental disorders (Buschmann et al. 2007). Main fluphenazine (FLU) metabolites in the human body are its sulfoxide and 7-hydroxide. Unmetabolized FLU and its two main metabolites account for 44 % of all FLU species that are excreted via the urine (Heyes and Robinson 1985). N-dealkylated FLU and FLU N-oxide were found in animal experiments but not in human urine. Further human metabolites include phase II reactions with glucuronation of hydroxyl groups (piperazine side chain and the FLU-7hydroxide; Heyes and Robinson 1985; Jackson et al. 1991) and cleavage of the piperazine ring to form an ethylendiamine metabolite (Breyer et al. 1974). FLU has a log kOW of 5.67 (Burgot and Burgot 1990) and due to its piperazine side chain two pka values at 3.9 and 8.1 (Sackett and McCreery 1979). Triflupromazine (Robinul®, Fig. 5—first line) was launched in 1957 for the treatment of schizophrenia but due to its antiemetic effects, it was also used in the management of

3163

nausea and vomiting (Buschmann et al. 2007). Main human metabolites result from mono- and dimethylation of the nitrogen atom of the aliphatic side chain as well as sulfoxidation of the phenothiazine core (Breyer et al. 1974). Triflupromazine (TPRO) has a log kOW of 5.17 (Franke et al. 1999) and a pka value of 9.2 (Sackett and McCreery 1979). Trifluoperazine (Stelazine®, Fig. 7—first line) was clinically introduced in 1959 as antipsychotic and anxiolytic drug. Due to its severe extrapyramidal side effects, its use nowadays has been declined in many parts of the world (Buschmann et al. 2007). Main metabolites include sulfoxidation, N-demethylation, 7-hydroxylation, as well as cleavage of the piperazine ring (Dachtler et al. 2000). The same ethylendiamine metabolite like in the case of FLU was found in human urine (Breyer et al. 1974). Trifluoperazine (TPER) has a log kOW of 5.10 (Franke et al. 1999) and like FLU due to its piperazine side chain, two pka values at 3.9 and 8.1 (Sackett and McCreery 1979). Polyfluorination Polyfluorinated compounds have become very popular in different areas of modern chemistry and are produced in high-scale quantities. In general, organic fluorinated pollutants are thought to be highly persistent against biodegradation and usually belong to the group of so-called persistent and bioaccumulative (P&B) chemicals. Therefore, concern has risen about their environmental fate (Lindstrom et al. 2011; Tressaud 2006). In a recent study, which listed 610 P&B commerce chemicals, 62 % were halogenated and 30 % fluorinated; in total, 181 compounds (Howard and Muir 2010). Besides being used in pharmacology, aromatic CF3 substitutions are also very popular in agrochemistry. Twentyeight percent of all halogenated agrochemicals in the USA, which were produced since 1940, are fluorinated (Jeschke 2004). In the late 1990s, it was found that more than 50 % of all fluorinated agrochemicals are trifluoromethylated aromatic compounds (Key et al. 1997). Fluorination of agrochemicals increases lifespan in the field as well as toxicity and bioactivity (Mueller 2000). Fluorination has become common in drug synthesis due to the fact that fluorinated drugs circulate longer in the body prior to phases I or II metabolism. The addition of fluorine moieties to drugs improves their stability against human metabolism and increases lipophility, basicity, and binding affinity to the target protein, therefore, extending dwell time in the target tissue (Bohm et al. 2004). Of all new registered pharmaceuticals in the USA in the year 2002, approximately 29 % were fluorinated (Bohm et al. 2004). Very often, aromatic drug compounds are substituted with CF3 groups. Popular examples are the selective serotonine reuptake inhibitor (SSRI) fluoxetine (Prozac®), the anti-inflammatory drug celecoxib and the platelet antiaggregant triflusal.

3164

Consumption and environmental fate Psychiatric drugs have become the most important prescription drugs worldwide, accounting for an economic volume of US$20 billion per year (Buschmann et al. 2007). However, the strong side effects of phenothiazine compounds have led to a decreasing utilization for some of them. Of all three investigated trifluoromethylated phenothiazine drugs, only FLU still was prescribed in Germany in the year 2010 (Schwabe and Paffrath 2010). 7.6 million defined daily doses (DDD) of oral FLU pills (each tablet accounting for 10 mg of pure FLU) and 2.9 million DDD of injected FLU depot formulations (each depot accounting for 1 mg of pure FLU), were prescribed (Schwabe and Paffrath 2010; WHO 2011). This corresponds to a total amount of 78.9 kg FLU only in that year in Germany. In a recent study, it was found that trifluoromethylated phenothiazine drugs are non high-production volume pharmaceuticals, which have properties to be classified as P&B chemicals (Howard and Muir 2011). Fluoxetine, one of the mostly prescribed SSRI antidepressants, was detected in surface waters (Kolpin et al. 2002) and found to be highly persistent and bioaccumulative (Howard and Muir 2011). So far, none of the compounds in this study has been detected in environmental samples even though they are classified as persistent (Howard and Muir 2011). Only recently, it was shown that biotic and abiotic degradation of chlorpromazine under environmental conditions was not complete resulting in the formation of numerous transformation products (TRPs; Trautwein and Kümmerer 2012). So far, no data on biodegradation of FLU, TPRO, and TPER have been reported. Considering abiotic elimination or transformation, e.g., by photolysis, some information is available for FLU (Miolo et al. 2006). Transformation and QSAR Depending on the environmental (e.g., surface water) or technical (e.g., sewage treatment plant; STP) compartment a pollutant enters, different transformation reactions can take place. Besides natural biotic (e.g., microbial degradation) or abiotic transformation (e.g., photolysis), technical processes like UV irradiation and ozonisation in drinking water facilities can be responsible in the formation of stable TRPs (Boxall 2009). Recently, it was shown that bacterial transformation of drugs can create entirely new dead-end TRPs, which are different to human metabolites and have not been reported before (Trautwein and Kümmerer 2011). The research about the environmental impact of these compounds has become a huge task (la Farre et al. 2008), especially because of lacking knowledge considering persistence, bioaccumulation, and toxicity of TRPs. Even though fluorination of organic compounds is generally thought to improve their persistence and bioaccumulation, there exist some mechanisms for defluorination, especially for

Environ Sci Pollut Res (2012) 19:3162–3177

mono- and trifluoromethylphenols. Besides reactions which only take place under laboratory conditions, there are also some processes which can occur in the free environment: fluorophenols were shown to be biodegradable by different Rhodococci species in wetlands (Boersma et al. 2001). The CF3-substituted aromatic drug fluoxetine was shown to be phototransformed upon UV irradiation (Lam et al. 2005). Environmental TRPs of drugs are usually only formed in low concentrations within complex matrices so that isolation and purification is very difficult. Further, these compounds are often not available commercially, which makes the performance of toxicity tests with pure substances impossible. Preparative chromatography might be a technique to isolate specific TRPs out of a huge mixture of compounds embedded in a complex matrix. However, recoveries of such procedures might be too little to perform many tests. Therefore, other tools for the environmental risk assessment of TRPs are needed. One possibility is the usage of hyphenated computer software packages, which are based on structure activity relationship (SAR) calculations. Once the structure of any TRP is elucidated via liquid chromatography–multiple stage mass spectrometry (LC-MSn) or nuclear magnetic resonance (NMR), these data can be used in quantitative structure activity relationship (QSAR) programs in order to predict different toxicological endpoints (Escher et al. 2009).

Materials and methods Biodegradation in the Closed Bottle test The Closed Bottle test (CBT) is used for aerobic biodegradation testing of organic compounds under low nutrient content and low bacterial density (101–104 colony forming units (CFU)/mL) therefore simulating conditions of a natural surface water body. Substances that pass the test are classified as readily biodegradable and therefore are expected not to reach or accumulate in the aquatic environment. The CBT was performed according to the Organisation for Economic Cooperation and Development (OECD) test guideline (OECD 301D 1992) at room temperature (20±1 °C) in the dark as described elsewhere in detail (Trautwein and Kümmerer 2011). The concentration in the test bottles was adjusted to 3.01 mg/L (FLU), 2.94 mg/L (TPRO), and 2.54 mg/L (TPER), respectively, corresponding to a ThODNH3 (theoretical oxygen demand without considering a possible nitrification) of 5 mg/L. All test bottles were inoculated with an aliquot of the effluent from the municipal STP in Kenzingen (Germany; 13,000 population equivalents). Two drops of inoculum (60 μL) were added to 1 L of medium, resulting in approximately 500 CFU/mL.

Environ Sci Pollut Res (2012) 19:3162–3177

Aerobic biodegradation was monitored by measuring the oxygen concentration in the test bottles with an optode oxygen sensor system (Fibox 3 PreSens, Regensburg, Germany), which is based on the physical principle of dynamic luminescence (Huber and Krause 2006; Wolfbeis 2002). In order to get kinetic information about the degradation process, oxygen and temperature were measured each day. The test guideline classifies a compound as “readily biodegradable” if biodegradability, expressed as a percentage of oxygen consumed in the test bottle (ThOD), exceeds 60 % within a period of 10 days after it reached 10 % ThOD. In addition to the test guideline, samples were taken at the beginning and at the end of each test and stored at −20 °C for later LC-MSn analysis. Photochemical transformation under indoor light The abiotic photolytic transformation was assessed with flasks of stock solutions of 50 mg/L. The solutions were prepared with deionised water in a 100 mL round bottom flask and stored in the dark until use. The samples were not exposed directly to sunlight but some daylight during preparation and handling of the solutions. Temperature conditions were room temperature at 20 °C. After 2 h of interior laboratory light exposure, samples were stored at −20 °C in the dark until LC-MSn analysis. Identification of transformation products by HPLC-MSn An high-performance liquid chromatography (HPLC) system (HPLC 1100 series, Agilent Technologies, Waldbronn, Germany) was used for chromatographic analysis. Separation was performed on a monolithic silica gel column with a characteristic bimodal pore structure (Chromolith Performance RP-18 endcapped 100–3, 2 μm; Merck Chemicals, Darmstadt, Germany) coupled to a precolumn (Chromolith Guard Cartridge RP-18e 5–3, 2 μm; Merck Chemicals). Eluents used were 10 mmol ammonium acetate in deionised water (solution A) and 100 mmol methanol LC–MS grade (solution B). The following linear gradient was applied: 0 min 20 % B, 25 min 40 % B, 30 min 52 % B, 43 min 65 % B, 44 min 95 % B, 45 min 95 % B, 45.01 min 20 % B. The sample injection volume was 5 μL and the flow rate was 1.0 mL/min. Post-run time was set to 2 min resulting in a total run time of 47.01 min. Standards (0.1, 1, 2.5, 5, and 10 mg/L, n03) were used to establish a linear calibration curve. Calibration curve parameters were r2 01.000 for FLU and TPRO and r2 00.996 for TPER. The precursor ion of FLU eluted at tr 036.2 min with a mass to charge ratio of m/z 438.2. TPRO was found at tr 037.5 min with m/z 353.1 and the precursor ion of TPER gave a peak at tr 037.5 min with m/z 408.2.

3165

The Agilent 1100 Series HPLC system was coupled to a Bruker Esquire 6000 plus ion trap mass spectrometer equipped with a Bruker data analysis system (Bruker Daltonik GmbH, Bremen, Germany). An atmospheric pressure interface with electrospray ionization was used for the generation of positive (+) charged molecular ions. Operating conditions of the ion generation were: −500 V end plate and −3,100 V capillary voltage relative to the needle, 50.00 psi (343 kPa) nebulizer pressure, 12.0 L/min nitrogen dry gas flow at a dry temperature of 365 °C, and capillary exit voltage of 114.9 V. The operating conditions of the ion transport and focusing were: 40 V skimmer (cone voltage), octopole radiofrequency amplitude of 150.0 Vpp and −60 V at lens 2. Parameters of the ion trap were set with the target mass of each phenothiazine parent compound respectively and a maximum accumulation time of 200 ms with a scan range from m/z 20 to 1,000. For sample measurements without fragmentation experiments, averages were set to six spectra with an accumulation time of 15 ms. For structural elucidation of TRPs, samples were analyzed by the auto MSn mode of the Esquire software, where ionized TRPs were isolated and fragmented up to MS4. Here, averages were set to one spectrum with an accumulation time of 75 ms and a precursor selection for further fragmentation of the three most abundant ions was chosen. Absolute threshold for MS2 was set to 5,000 ion counts for n>2 to 20 ion counts. High frequent isolation and fragmentation procedures of the Esquire 6000 plus apparatus brought along some mass shifts during measurements. Therefore, the m/z values of some precursor and product ions differ from the exact mass up to m/z ±0.3. However, as these inaccuracies are the same for all target compounds, the obtained data could still be used for structural elucidation. Obtained fragment ions were compared as far as possible with literature data. Molecular structures of typical phenothiazine fragment ions after electrospray ionization have been identified for chlorpromazine (Smyth 2003; Trautwein and Kümmerer 2012) and piperazine substituted compounds (Hubert-Roux et al. 2010; Smyth 2005). Especially the fragmentation of aromatic amine side chains was interesting since it was shown that cleavage of C–N bounds leads to unstable carbonium ions RCH-CH2+ which readily react further to stable alkenes RCH0CH2 (Joyce et al. 2004). Chemicals All the chemicals used as nutrients, for rinsing glassware and for DOC analysis, were at least of >98.5 % purity. Methanol used for chromatographic elution was of gradient grade (Chromasolv® 99.9 %, Sigma-Aldrich, Fluka Analytical, Darmstadt). Fluphenazine (CAS 146-56-5), triflupromazine (CAS 440-17-5), and trifluoperazine (CAS 1098-60-8) were

3166

Environ Sci Pollut Res (2012) 19:3162–3177

Fig.1 Aerobic biodegradation of fluphenazine (FLU) in the Closed Bottle test

100 90

biodegradation [%]

80 Sodium acetate (n=2) FLU toxicity control (n=2) FLU toxicity control computed FLU (n=2 )

70 60 50 40 30 20 10 0 -10

0

2

4

6

8

10 12 14 16 18 20 22 24 26 28

time [days]

purchased from Sigma-Aldrich (Munich, Germany) with 98 % purity or higher. QSAR For the identification of potential environmental risks deriving from phenothiazines and their TRPs, two different QSAR tools were used: the EPI Suite software (EPIWEB 4.1) from the U.S. Environmental Protection Agency (US EPA 2004) and the Computer Automated Structure Evaluation (CASE) Ultra software (Version 1.4.4.6, MultiCASE Inc., Beachwood, USA). Simplified molecular input line entry specification (SMILES) codes from the molecular TRP structures were taken as input. Since CASE Ultra splits the molecule up into fragments, which then are compared with a database (Klopman et al. 2005), it can happen that unknown fragments appear and therefore the applicability is out of domain (OD). Further, it is possible that some fragments of a structure have activating properties whereas others are inactivating. In such cases, the final result is 100,0%

100%

LC/MS drug recovery

Fig. 2 Analytical recovery of fluphenazine (FLU), triflupromazine (TPRO) and trifluoperazine (TPER) after 28 days incubation in the Closed Bottle test (CBT); Tox toxicity control bottles

inconclusive. Obtained CASE units classify a substance as inactive (10–19 CASE units), marginal (20–29), active (30– 39), very active (40–49), or extremely active (50–99). Physicochemical parameters that were predicted with both software packages are water/octanol partition coefficient log kOW and water solubility SH2O. Organic matter soil/water partition coefficient log KOC, bioconcentration factor log BCF, and Henry law constant LC were only calculated with EPIWEB 4.1 since the CASE Ultra software offers no calculation of this endpoint. Further parameters, which were predicted, are elimination (only EPIWEB 4.1) and biodegradation in STP and ready biodegradability in the Ministry of International Trade and Industry (MITI) test, which is not directly comparable to the Closed Bottle test. The MITI test often indicates better biodegradability than the CBT because of higher bacterial density and diversity as well. The obtained results were compared with experimentally determined values and with critical limits according to EU and US guidelines (European Commission 2003; US EPA 1999).

93,7%94,1% 95,1% 95,1%

99,1% 98,7% 92,4% 93,2%

90,8% 89,3%

86,3% 84,2%

80%

60%

40%

20%

0% d -0

8d -2 2 x 8d To -2 ER 1 x TP To 8d ER -2 TP 2 ER 28d TP 1 ER TP 8d -2 2 x 8d To -2 O R 1 TP Tox O R 8d -2 TP 2 O d R 28 TP 1 O R TP 8d -2 2 x 8d To U -2 1 FL x To U 8d FL -2 2 U d FL - 28 1

BT

U FL

C

Environ Sci Pollut Res (2012) 19:3162–3177

3167

Results Biodegradation in the Closed Bottle test The CBT was valid according criteria of the OECD test guideline (OECD 301D 1992) since 84±3.5 % ThOD of the quality control substrate (sodium acetate) were degraded within a 10-day window (>60 % ThOD are necessary; Fig. 1, Electronic supplementary material (ESM) 1 and 2) and oxygen concentrations in all bottles did not fall below

0.5 mg/L at any time. Oxygen depletion in the blanks after 28 days was 0.58 mg/L and therefore less than 1.5 mg/L. The biodegradation values determined in the CBT by monitoring the oxygen concentration were FLU0–2.5± 0.2 %, TPRO02.0±0.8 %, and TPER0−5.4±4.7 % (average±range; n02) at test end after 28 days. These values characterize all three compounds as being not readily biodegradable. In the toxicity control bottles, the following biodegradation values were calculated based on the composition of the mixture of test compound and quality

Table 1 Chromatographic and mass spectrometric (MS) data of the nine identified transformation products (TRP) in the CBT and indoor light photodegradation test Transformation product

TRP (m/z)

RT (min)

MS/MS (m/z) (precursor in bold)

MS3 (m/z) (precursor in bold)

MS4 (m/z) (precursor in bold)

Fluphenazine FLUTRP (S/N>55,000)

1

414.2

16.1

414.4 >143.1, 171.1, 256.2

2

470.4

18.6

470.4>280.7, 324.0, 353.4, 147.2

143.1>70.2, 99.2, 113.1 171.1>70.3, 98.9, 143.0 256.0>211.5, 224.0, 238.4, 240.1 280.7>210.9, 248.0 324.0>267.0, 280.1, 295.9, 306.9 353.4>266.2, 310.2

3

454.2

20.0

454.5>171.1, 226.1, 280.2, 324.2, 366.2

226.1>150.9 280.2>210.9, 248.0

99.2>70.3 238.4>210.0 240.1>210.0 295.9>263.1, 277.1 306.9>266.1 266.2>238.9 310.2>293.8, 266.0 296.0>226.9, 236.1 306.6>265.9

324.2>133.2, 265.9, 296.0, 306.6 366.2>208.0, 226.1, 254.1 239.3>207.1 256.4>209.8, 211.9, 224.0, 238.6 266.0>238.9 284.2>212.8, 239.1, 251.9, 239.0, 256.2

226.1>197.9

267.1>222.0 291.0>192.9, 234.9, 262.8 324.2>266.1, 275.9, 307.0 338.4>266.4, 280.0, 306.6 324.3>307.0, 229.9, 200.9, 291.0, 296.0

307.0>266.0 291.7>197.8, 262.1 306.6>262.0, 271.1

113.2>70.2 141.2>70.3, 113.1 238.9>206.3 256.0>211.0, 223.9, 238.4 284.1>225.2, 266.1 266.7>246.9

238.4>210.2

280.2>248.0 324.0>266.8, 280.0, 296.4, 307.2 353.2>265.9, 310.4 367.0>265.9, 284.1, 310.0 141.1>70.3, 113.1 295.5>268.0 316.3>230.3, 272.7, 288.0, 300.9 324.2>266.5, 290.6, 296.1, 307.2

280.2>247.9 296.4>262.6, 276.9 307.2>265.9 310.4>265.9, 294.7 268.0>248.9 288.0>245.0 290.6>194.1 296.1>277.0 307.2>266.0

Fluphenazine FLUTRP (S/N>35,000)

Fluphenazine FLUTRP (S/N>15,000)

Triflupromazine TPROTRP (S/N>25,000)

Triflupromazine TPROTRP (S/N>35,000)

1

329.2

11.5

329.3>239.3, 256.4, 266.0, 284.2

2

369.2

16.2

369.4>266.0, 267.1, 291.0, 296.1, 324.2, 338.4

385.2

18.4

384.2

18.3

385.3>324.3, 354.0, 296.1, 267.0, 307.1 384.4>113.1, 141.1, 238.9, 256.0, 284.1

Triflupromazine TPROTRP 3 (S/N>4,000) Trifluoperazine TPERTRP 1 (S/N>30,000)

Trifluoperazine TPERTRP (S/N>15,000)

Trifluoperazine TPERTRP (S/N>8,000)

2

3

440.4

424.5

S/N signal/noise ratio, RT retention time

19.3

21.3

440.4>266.7, 280.2, 324.0, 353.2, 367.0

424.5>141.1, 295.5, 316.3, 324.2, 376.2

238.9>210.1 239.0>206.2 251.9>223.9 256.0>210.1

256.1>224.0 266.1>234.0, 237.9 266.8>198.0

3168

Environ Sci Pollut Res (2012) 19:3162–3177

control: FLU 031.6 ± 5.6 %, TPRO 034.8 ± 5.7 %, and TPER 38.0 ± 0.7 %. These values are all >25 % and therefore indicate no toxicity according to the test guideline (OECD 301D 1992). Analysis by LC-MSn of samples taken at days 0 and 28 demonstrated little removal of test compounds (Fig. 2) and showed new peaks in the chromatograms. Maximum FLU elimination was found in bottle FLU 1 (6.3 %). Maximum TPRO elimination was found in bottle TPRO Tox 1 (7.6 %) and highest TPER elimination took place in bottle TPER Tox 2 (15.8 %). Peak data were checked with the photolysis products found in “Formation of photoproducts under indoor light” section and matched with m/z values and retention times. Comparison showed that the elimination happened by photochemical transformation. Formation of photoproducts under indoor light LC-MSn analysis of stock solutions exposed 2 h to indoor light showed the formation of three TRPs for each phenothiazine. They eluted at the same retention times and with the same m/z values as the TRPs of the corresponding CBT samples, respectively. Molecular structures of these TRPs were elucidated according to “Identified transformation products by HPLC-MSn” section

In order to elucidate the molecular structures of the photoproducts formed in the CBT and the photolytic test, fragmentation spectra of higher order were acquired using the Auto MSn mode of the mass spectrometer software. Depending on peak intensity and fragmentation characteristics of each TRP, up to MS4 spectra were generated (Table 1). The obtained mass spectra and fragment ions were used to draw fragmentation schemes and propose structures for the photoproducts. All TRPs eluted at earlier retention times than their parent compound, respectively, which indicates higher polarity of the TRPs compared to the respective parent compound. MSn experiments were performed according to “Identified transformation product by HPLC-MSn” section. All samples from the CBT as well as samples of the photolytic experiment were measured with the same LC-MSn method. For FLU, the most intense TRP peak was found at tr 016.1 min with an S/N>55,000 (Table 1, Fig. 3). In the Auto MSn mode, FLUTRP 1 0m/z 414.2 gave several product ions (Table 1), which were fragmented again so that a complex fragmentation pattern of mass spectra emerged (ESM 3). These MS2–4 data were used for structural elucidation. MS2 fragments at m/z 171 and m/z 143 allowed for to rule out oxidation at the piperazine side chain. Further fragments

Intens. x108 1.5

m/z 438.2

FLU and photoproducts - roomlight 2h (TIC)

m/z 470.2

0.5

m/z 454.2

1.0 m/z 414.2

Fig. 3 Total ion currents (TIC) of fluphenazine (FLU) and its transformation products (TRPs) after 2 h indoor light irradiation in aqueous solution (first line) and after 28 days bacterial incubation in the Closed Bottle test (second and third line). 1, m/z 414.2; 2, m/z 470.2; 3, m/z 454.2; and 4, m/z 438.2 (FLU)

Identified transformation products by HPLC-MSn

0.0 Intens. x107

FLU and TRPs - CBT 28d test bottles (TIC) 4

3

2

1 1

2

0 Intens. x107

3 FLU and TRPs - CBT 28d toxicity control bottles (TIC) 4

3

2

1 1 0 5

10

15

2

3 20

25

30

35

40

Time [min]

Environ Sci Pollut Res (2012) 19:3162–3177

of m/z 414 indicated this TRP to be 2-carboxyphenazine (Fig. 4). FLUTRP 2 0m/z 470.2 was the second most intensive TRP and eluted at tr 018.6 min with an S/N>35,000 (Table 1, Fig. 3). MS2 fragments at m/z 147 and m/z 324 indicated oxidation at the side chain and the phenothiazine core. Further characteristic MS2–4 product ions of m/z 470.2 could be attributed to sulfoxidized exocyclic FLU-N-oxide (ESM 4 and 5). FLUTRP 3 0m/z 454.2 was the third TRP and gave peaks at tr 020.0 min with an S/N>15,000 (Table 1, Fig. 3). MS2 fragments at m/z 171 excluded oxidation at the piperazine side chain and further MS2–4 data indicated m/z 454.2 to be FLU sulfoxide (ESM 6 and 7). For TPRO, the most intense TRP peak was found at tr 0 16.2 min with an S/N>35,000 (Table 1, Fig. 5). The Auto MS2–4 fragmentation of TPROTRP 2 0m/z 369.2 gave several product ions (Table 1 and ESM 8). MS2 fragments at m/z 338 indicated extrusion of a SO moiety (48 Da) and further MS2–4 product ions indicated m/z 369.2 to be sulfoxidized TPRO (Fig. 6). TPROTRP 1 0m/z 329.2 was the second most intense TRP and eluted at tr 011.5 min with an S/N>25,000 (Table 1,

3169

Fig. 5). The obtained characteristic MS2–4 fragment ions indicated m/z 329.2 to be 2-carboxypromazine (ESM 9 and 10). TPROTRP 3 0m/z 385.2 was the third TRP (S/N>4,000) and gave in all samples peaks at tr 018.4 min (Table 1, Fig. 5). MS2−4 data indicated m/z 385.2 to be an exocyclic N-oxidized TPRO sulfoxide (ESM 11 and 12). For TPER, the most intense TRP peak was found at tr 019.3 min with an S/N > 30,000 (Table 1, Fig. 7). Auto MSn fragmentation of TPERTRP 2 0m/z 440.2 gave several product ions (Table 1 and ESM 13). The acquired MS2–4 mass spectra and fragmentation pattern of m/z 440.2 indicated this TRP to be sulfoxidized exocyclic TPER-N-oxide (Fig. 8). TPERTRP 1 0m/z 384.2 was the second most intense TRP and eluted at tr 018.3 min with an S/N>25,000 (Table 1, Fig. 7). MS2 fragments at m/z 141 and m/z 113 excluded oxidation at the piperazine side chain. Further, MS2–4 data of 384.2m/z could be attributed to 2-carboxyperazine (ESM 14 and 15). TPERTRP 3 0m/z 424.2 was the third TRP and gave in all samples peaks at tr 021.3 min with an S/N>8,000 (Table 1, Fig. 7). MS2 fragments at m/z 141

Fig.4 Fragmentation scheme proposal for the most intense photolytic transformation product of fluphenazine: 2-carboxyphenazine (m/z 414.2 at tr 016.1 min) according to the acquired MS2–4 spectra

3170 Intens. x108

m/z 353.1

TPRO and photoproducts - roomlight 2h (TIC)

2.0

0.5

m/z 385.2

1.0

m/z 369.2

1.5 m/z 329.2

Fig. 5 Total ion currents (TIC) of triflupromazine (TPRO) and its transformation products (TRPs) after 2 h indoor light irradiation in aqueous solution (first line) and after 28 days bacterial incubation in the Closed Bottle test (second and third line). 1, m/z 329.2; 2, m/z 369.2; 3, m/z 385.2; and 4, m/z 353.1 (TPRO)

Environ Sci Pollut Res (2012) 19:3162–3177

0.0 Intens. x107 4

TPRO and TRPs - CBT 28d test bottles (TIC) 4

3 2 1 2

1

0 Intens. x107

3 TPRO and TRPs - CBT 28d toxicity control bottles (TIC) 4

3 2 1 2

1

0 0

5

10

15

3 20

25

30

35

40

Time [min]

and m/z 113 excluded oxidation of the side chain and indicated extrusion of a SO moiety. Together with further MS2–4 data, m/z 424.2 was tentatively identified as TPER sulfoxide (ESM 16 and 17).

the unchanged drugs and their carboxylated TRPs (Table 2, column 12). Ready biodegradation cannot be expected for any substance (Table 2, column 13).

Predicted QSAR parameters

Discussion

Table 2 shows the predicted environmental parameters with the EPI and CASE software. Most calculations are based on the log kow value of each compound (Table 2, column 6), which was estimated with the corresponding SMILES codes of each structure (Table 2, column 5). Water solubility at 25 °C (Table 2, column 7) shows highest values for S- and N-oxidized compounds (0.05–94.57 mg/L), whereas solubilities of unchanged phenothiazines are low, ranging from 0.001 to 0.27 mg/L. Strong solubility differences between the EPI and CASE results were observed. Results of log Koc and log BCF prediction (Table 2, columns 8 and 9) show that the unpolar parent compounds have approximately a 20-fold stronger tendency for soil adsorption and bioaccumulation than their TRPs. The calculated Henry constants show for all compounds very low values ranging from 4.32×10−9 to 3.08×10−28 atm×L/mol (Table 2, column 10). Finally, quantitative prediction of elimination in a STP show negligible STP biodegradation values for all compounds ranging from 0.1 to 0.8 % (EPI Suite) and mostly inconclusive CASE results (Table 2, column 11), besides a low biodegradability of carboxypromazine (39.2 CASE units). EPI results for sludge adsorption show high abiotic elimination rates for

Biodegradation Tricyclic phenothiazine drugs are pharmaceuticals, which have been used now for more than half a century. However, for fluorinated derivates, no data about biodegradation was available in literature so far and the physicochemical properties of these compounds give reason of concern about their environmental fate. Only recently, a comprehensive study has shown that the most popular chlorinated phenothiazine drug chlorpromazine cannot be environmentally degraded and revealed the formation of numerous biotic and abiotic dead-end transformation products (Trautwein and Kümmerer 2012). These findings correlate well with the results of this study, as it was shown here that trifluoromethylated phenothiazine drugs have to be classified as not readily biodegradable according to OECD guidelines (OECD 301D 1992). Toxicity against bacteria of the inoculum was not found in respect to the test guideline, since more than 50 % of the control substrate sodium acetate in the toxicity control bottles was biodegraded within 2 days and approximately 70 % at test end (Fig. 1, ESM 1 and 2). However, it should be considered that bactericide properties of phenothiazines are more of a bacteriostatic origin (Wesołowska et al. 2009)

Environ Sci Pollut Res (2012) 19:3162–3177

3171

Fig. 6 Fragmentation scheme proposal for the most intense photolytic transformation product of triflupromazine: TPRO sulfoxide (m/z 369.2 at tr 0 16.2 min) according to the acquired MS2–4 spectra

which might result in some lag phase before any effect can be observed. A comparison between the computed and measured toxicity curves supports this postulation, as it shows that after 5 days, the curves start to differ (Fig. 1, ESM 1 and 2) with lower measured than expected computed biodegradation values. Hyphenated LC-MSn analysis revealed at test the formation of three TRPs for each phenothiazine (Figs. 3, 5, and 7). However, the fact that as the test begins, residual concentrations of these TRPs were also detected, questioned the consideration that these TRPs were of bacterial origin. Therefore, another experimental setup with interior indoor light irradiated stock solutions of all three compounds showed all identified TRPs to be of abiotic origin. Phototransformation The formation of abiotic phenothiazine TRPs upon irradiation with UV lamps has been reported before in literature (Miolo et al. 2006; Motten et

al. 1985), but so far it was shown in one case only that even under interior indoor light conditions photochemical TRPs can easily form (Trautwein and Kümmerer 2012). Within trifluoromethylated phenothiazine drugs, only FLU has been investigated so far considering its photolytic degradation. The transformation of FLU to 2-carboxyphenazine and exocyclic FLU N-oxide was found under irradiation with ultraviolet A (UVA) light (λmax 0365 nm; Miolo et al. 2006) whereas the photochemical formation of FLU sulfoxide has not been reported before. N-oxidized FLU could be photolytic precursor of N-oxidized FLU sulfoxide. Since the experiments in this study used only diffuse indoor light and no active UV lamp, differences in excitation wavelengths and irradiation energies could explain these different transformation behaviors. The identification of carboxylated and S-/N-oxidized photoproducts of TPER and TPRO was shown for the first time in this study. It seems that the photochemical

3172 Intens. x108

m/z 408.2

TPER and photoproducts - roomlight 2h (TIC)

2.5 2.0

1.0 0.5

m/z 424.2

1.5 m/z 384.2 m/z 440.2

Fig. 7 Total ion currents (TIC) of trifluoperazine (TPER) and its transformation products (TRPs) after 2 h indoor light irradiation in aqueous solution (first line) and after 28 days bacterial incubation in the Closed Bottle test (second and third line). 1, m/z 384.2; 2, m/z 440.2; 3, m/z 424.2; and 4, m/z 408.2 (TPER)

Environ Sci Pollut Res (2012) 19:3162–3177

0.0 Intens. x107

TPER and TRPs - CBT 28d test bottles (TIC) 4

3

2

1 1 2 3

0 Intens. x107

TPER and TRPs - CBT 28d toxicity control bottles (TIC) 4

3

2

1 1 2 3

0 0

5

10

transformation of an aromatic CF3 substituent to its carboxylic acid is a common mechanism for trifluoromethylated aromatic compounds, as this was also reported for the antidepressant fluoxetine (Lam et al. 2005) and the antiplatelet antiaggregant drug triflusal (Bosca et al. 2001). Interestingly, all three phenothiazine compounds showed similar transformation behavior. The photolytic formation of phenothiazine sulfoxides was shown to be dependent on the formation of a light-induced sulfuric peroxy radical intermediate (Motten et al. 1985) whereas for the other types of TRPs yet no photochemical transformation rule has been established. The fact that only exocyclic but no endocyclic N-oxides were identified shows the stability of the endocylic phenothiazine nitrogen against oxidation. Even though all structures were elucidated solely by LC-MSn techniques, this approach showed to be of a high reliability. This was the case since several literature data with similar experiments and results were available for comparison. Some of these studies proofed their obtained mass spectrometric results by NMR spectra, which is the gold standard for structural elucidation of low-molecular unknown compounds. Phototoxicity Photosensitization is one of the most popular side effects of phenothiazine drugs (Viola and Dall’Acqua

15

20

25

30

35

40

Time [min]

2006). Especially chlorinated phenothiazines like chlorpromazine and perphenazine were found to have strong phototoxic properties (Gocke 1996). Several studies have identified different mechanisms including the formation of radicals (Ciulla et al. 1986) and/or dimers (Kochevar and Hom 1983). However, for a long time, it was thought that photolytic cleavage of CF3 substituents and following covalent binding with peptides and toxic reactions is unlikely. The results of this study show that at neutral pH even under interior indoor light, hydrolytic photolytic cleavage transforms the trifluoromethyl into a carboxylic moiety. The photochemical transformation of the CF3 to a COOH moiety was found to be one of the major reasons for phototoxicity of FLU since upon photolytic activation not only water but also proteins like calmodulin can covalently bind with a peptide bound and therefore form allergenic photoadducts (Caffieri et al. 2007). Caffieri et al. (2007) found also that amino acids with another nucleophilic moiety (e.g., lysine and tyrosine) can even connect two FLU molecules and form a FLU diadduct. This mechanism was also shown for glutathione. Caffieri et al. (2007) further postulated that these photoadducts might act as antigen and be the reason for the well-known phototoxic side effects of FLU like lifethreatening agranulocytosis. Other phototoxic effects, which were shown for FLU under UVA irradiation, were hemolysis

Environ Sci Pollut Res (2012) 19:3162–3177

3173

Fig. 8 Fragmentation scheme proposal for the most intense photolytic transformation product of trifluoperazine: exocyclic N-oxidized TPER sulfoxide (m/z 440.2 at tr 019.3 min) according to the acquired MS2–4 spectra

of mouse erythrocytes and toxicity on cultured murine fibroblasts (Elisei et al. 2002). It was also found that mitochondria and plasma membranes are targets of UVA-induced FLU toxicity (Bastianon et al. 2005). Even photogenotoxic effects have been reported for FLU as irradiation of supercoiled plasmid DNA led to single strand break where not intercalation but externally bound FLU molecules seemed to be the reason (Viola et al. 2003). Reported phototoxicity data for the two other phenothiazine compounds is rare. TPER was shown to inhibit calmodulin-activated enzymes and that binding to calmodulin can be made irreversible by irradiation with UV light (Prozialeck et al. 1981). Prozialeck et al. (1981) could not explain this mechanism but postulated the formation of a covalent bound between TPER and calmodulin. These findings match with the results of Caffieri et al. (2007). Like FLU, photohemolysis of human erythrocytes was also reported for TPER and TPRO (Eberlein-Koenig et al. 1997). Environmental risks Eventhough antipsychotic drugs are administrated in very low doses, which result in minimal PEC values, their critical effect concentrations are also very low so

that potential environmental risks should in any case be taken into consideration. Howard and Muir (2011) found that trifluoromethylated phenothiazine drugs have to be classified as P&B chemicals. FLU and TPER were included in the HSDB drug list (HSDB 2011), which shows their potential to be hazardous chemicals. So far, these compounds have not been detected in environmental samples even though they are classified as persistent (Howard and Muir 2011), which arise questions upon the degradation behavior of these drugs. This study showed for the first time that trifluoromethylated phenothiazine drugs are not readily biodegradable. However, upon light exposure, three different types of photoproducts can easily be formed: sulfoxides, N-oxides, and carboxylated derivates. As these photoproducts were also formed during the course of the CBT, it can be assumed that they are like their parent compounds not readily biodegradable. For a total of nine transformation products and their three parent compounds, environmental risk parameters were calculated (Table 2). In both software packages showing all unchanged phenothiazines, most carboxylated and some sulfoxidized TRPs showed log kow values>3 which

20.0 453.5 OCCN1CCN(CC1)CCCN(C20C3C0CC(C(F)(F)F)0C2) C40CC0CC0C4S30O

Sulfoxide

18.4 384.4 C[N+](C)([O-])CCCN(C10C2C0CC(C(F)(F)F)0 C1)C30CC0CC0C3S20O

Sulfoxidized exocyclic N-oxide

5.76 0.02 2.04 0.07

0.09 0.002 0.10 0.04

0.11 0.001 23.68 0.10 5.95 0.05 21.35 0.17

0.27 0.01 0.48 0.15 26.12 0.08 93.57 0.26

2.29 1.34 2.19E-020

2.72 1.84 2.73E-013

2.48 0.50 1.00E-14

3.86 3.32 4.32E-009

2.00 1.06 6.35E-024

2.42 1.57 7.92E-017

2.19 0.50 2.91E-018

3.51 2.99 1.26E-012

3.1

9.1

52.3

88.0

2.3

4.7

31.1

78.0

1.8

2.1

1.47 0.67 3.84E-021 1.04 0.17 3.08E-028

6.0

48.5

1.24 0.50 1.41E-022

2.73 2.29 6.09E-017

0.8 IC 0.5 39.2 CASE (scale 10–99) 0.2 IC 0.1 IC

0.7 IC 0.3 IC IC 0.1 IC 0.1 IC

0.5 UF/OD 0.1 IC 0.1 UF/OD 0.1 UF/OD

No UF/OD No UF/OD

No 0% No 0%

No 0% No 0% No UF/OD No UF/OD

No IC No IC No IC No IC

log log Henry LC STP/sludge elimination Ready KOC BCF (atm× biodegradation L/mol) Adsorption Biodegradation (MITI) sludge (%) (%)

FLU fluphenazine, TPER trifluoperazine, TPRO triflupromazine, RT retention time, MW molecular weight, SMILES simplified molecular-input line-entry specification, S solubility, BCF bioconcentration factor, STP sewage treatment plant, UF unknown fragments, OD out of domain, IC inconclusive, CASE computer-automated structure evaluation, MITI Ministry of International Trade and Industry Japan

16.2 368.4 CN(C)CCCN(C10C2C0CC(C(F)(F)F)0C1)C30CC0 CC0C3S20O

Sulfoxide

Carboxypromazine 11.5 328.4 CN(C)CCCN1C20CC0CC0C2SC30C1C0C(C(O)0O)C0C3

3.30 3.25 2.53 2.90

5.52 4.83 4.44 3.29

32.8 352.4 FC(F)(F)C10CC20C(C0C1)SC30CC0CC0 C3N2CCCN(C

19.3 439.5 CN1CC[N+](CC1)([O-])CCCN(C20C3C0 CC(C(F)(F)F)0C2)C40CC0CC0C4S3 0O

21.3 423.5 CN1CCN(CC1)CCCN(C20C3C0CC(CF)(F)F)0 C2)C40CC0CC0C4S30O

Sulfoxide

Sulfoxidized exocyclic N-oxide TPRO Unchanged

18.3 383.5 O0C(O)C10CC20C(C0C1)SC30CC0CC0C3N2CC CN4CCN(C)CC4

Carboxyperazine

5.11 4.79 4.03 3.24 2.88 3.20 2.12 2.85

4.13 4.13 3.05 2.65 1.91 2.61 1.14 2.25

log SH2O kOW 25 °C (mg/L)

37.5 407.5 FC(F)(F)C10CC20C(C0C1)SC30CC0CC0C3N2CCC N4CCN(C)CC4

18.6 469.5 FC(F)(F)C10CC(N(CCC[N+](CC2)([O-])CCN2CCO) C30CC0CC0C3S40O)0 C4C0C1

16.1 413.5 OC(C10CC(N(CCCN(CC2)CCN2CCO)C30CC0CC0 C3S4)0C4C0C1)0O

Carboxyphenazine

SMILES (g/mol)

36.2 437.5 OCCN1CCN(CC1)CCCN2C30CC0CC0C3SC40 C2C0C(C(F)(F)F)C0C4

RT MW (min)

Unchanged

Sulfoxidized exocyclic N-oxide TPER Unchanged

FLU

Compound

Table 2 Predicted environmental parameters of fluorinated phenothiazine drugs and their phototransformation products calculated with EPIWEB 4.1 (upper values in normal script) and CASE Ultra 1.4.4.6 (lower values in italic)

3174 Environ Sci Pollut Res (2012) 19:3162–3177

Environ Sci Pollut Res (2012) 19:3162–3177

marks them as potentially bioaccumulative (US EPA 1999). Experimentally determined log kow values (see “Introduction” section) for TPRO (5.17) and TPER (5.10) are between the two QSAR results, whereas the experimental log kow of FLU (5.67) is approximately 1.5 units higher than the EPI and CASE results. The calculated values for water solubility are low with strong differences between the two software results, especially considering the multiple oxidized TRPs. Whereas EPI finds S/N-oxidized TRP solubilities ranging from 2.04 mg/L (TPRO) to 93.57 mg/L (FLU), CASE results range from 0.07 mg/L (TPRO) to 0.26 mg/L (FLU). High chromatographic retention times of the S/N-oxidized TRPs indicate less polarity than carboxylated TRPs so that the CASE results seem to be more likely. Even though S- and N-oxides are two polar moieties, their dipole moments could be neutralized by each other. Considering the minimal PEC values which are in the nanogram per liter range, solubility values classify all TRPs as high mobile in the aquatic environment. EPI bioconcentration factors (log BCF values) range from 0.17 to 1.84 so that accumulation in tissues is unlikely. The calculated EPI log Koc values for all TRPs range from 1.04 to 2.72 so that no adsorption to soil should be expected. All EPI Suite Henry LC values are negligible low so that elimination through evaporation into the atmosphere is insignificant. Entering a sewage treatment plant, no biodegradation can be expected even though EPI values for sludge adsorption show that some abiotic elimination of the carboxylic TRPs might occur. Finally, values around zero for ready biodegradation prediction correspond with the experimental findings in this study which showed neither biodegradation for trifluoromethylated phenothiazine drugs nor their photochemical TRPs. Inconclusive CASE values for sludge and ready biodegradation show that the molecular structures of the TRPs are quite complex and neither this nor related ones are in the database. As data for biodegradability of phenothiazines were not available until now, it remains open how the coverage (in other words whether they are within the applicability domain) of EPI is for these types of chemical structures. Further, the detection of several unknown fragments, which are based on common environmental transformation pathways like sulfur and nitrogen oxidation, prove the lack of reference data and strongly indicate further research upon these compounds. The indication of applicability domains (OD) is an advantage of the CASE software as it not only calculates, but also evaluates data. Interestingly, oxidation at the sulfur atom of phenothiazine drugs is also a common pathway in human phase I metabolism so that S-oxidized FLU (Heyes and Robinson 1985), TPER (Dachtler et al. 2000), and TPRO (Breyer et al. 1974) are main metabolites in human excretion. The findings of this study now show that formation of trifluoromethylated phenothiazine S-oxides can occur enzymatically via cyctochrome P450 monooxygenases, as well as photochemically upon UV light exposure.

3175

Conclusion Three trifluoromethylated phenothiazine drugs were shown to be not ready biodegradable according to OECD 301D guidelines. However for each compound, three abiotic TRPs were identified which formed quickly in aerobic aqueous solutions upon light exposure and which were stable against biodegradation in the CBT. The calculation of some environmental parameters showed high mobility and persistence with some tendency for bioaccumulation. Further risk assessment on these drugs and their TRPs is therefore strongly recommended as phototoxicity was already reported in literature. For further toxicological investigations and measuring of environmental samples, not only phenothiazine parent compounds, but also their sulfoxides should be considered since their expected environmental concentrations is higher. Acknowledgments The authors wish to thank Georg Fuchs (Institute for Biology II, Microbiology, University Freiburg) for fruitful discussions about bacterial metabolism of trifluoromethylated aromatic compounds. The authors wish to thank especially Christoph Lederer (Leuphana University Lüneburg) for calculation of QSAR parameters with the CASE Ultra software. This work was supported through Deutsche Bundesstiftung Umwelt (DBU) through a scholarship for Christoph Trautwein (grant no. 20007/940).

References Bastianon C, Zanoni R, Miolo G et al (2005) Mitochondria and plasma membrane as targets of UVA-induced toxicity of neuroleptic drugs fluphenazine, perphenazine and thioridazine. Int J Biochem Cell Biol 37:901–908. doi:10.1016/j.biocel.2004.11.016 Boersma M, Solyanikova I, Van Berkel W et al (2001) F-19 NMR metabolomics for the elucidation of microbial degradation pathways of fluorophenols. J Ind Microbiol Biotechnol 26:22–34. doi:10.1038/sj.jim.7000027 Bohm H, Banner D, Bendels S et al (2004) Fluorine in medicinal chemistry. ChemBioChem 5:637–643. doi:10.1002/cbic.200301023 Bosca F, Cuquerella M, Marin M, Miranda M (2001) Photochemistry of 2-hydroxy-4-trifluoromethylbenzoic acid, major metabolite of the photosensitizing platelet antiaggregant drug triflusal. Photochem Photobiol 73:463–468. doi:10.1562/0031-8655(2001) 0730463POHTAM2.0.CO2 Boxall A (2009) Transformation products of synthetic chemicals in the environment, 1st edn. Springer, New York Breyer U, Gaertner HJ, Prox A (1974) Formation of identical metabolites from piperazine- and dimethylamino-substituted phenothiazine drugs in man, rat and dog. Biochem Pharmacol 23:313–322. doi:10.1016/0006-2952(74)90422-5 Burgot G, Burgot J (1990) Water/n-octanol thermodynamic parameters of some phenothiazines. Int J Pharm 62:R5–R7. doi:10.1016/ 0378-5173(90)90242-V Buschmann H, Holenz J, Párraga A et al (2007) Antidepressants, antipsychotics, anxiolytics: from chemistry and pharmacology to clinical application, 1. Wiley, Weinheim Caffieri S, Miolo G, Seraglia R et al (2007) Photoaddition of fluphenazine to nucleophiles in peptides and proteins. possible cause of immune side effects. Chem Res Toxicol 20:1470–1476. doi:10.1021/ tx700123u

3176

Ciulla TA, Epling GA, Kochevar IE (1986) Photoaddition of chlorpromazine to guanosine-5′-monophosphate. Photochem Photobiol 43:607–613. doi:10.1111/j.1751-1097.1986.tb05635.x Dachtler M, Händel H, Glaser T et al (2000) Identification of trifluoperazine metabolites using HPLC-NMR of test mixtures. Magn Reson Chem 38:951–956. doi:10.1002/1097-458X(200011) 38:113.0.CO;2-6 Eberlein-Koenig B, Bindl A, Przybilla B (1997) Phototoxic properties of neuroleptic drugs. Dermatology 194:131–135. doi:10.1159/ 000246080 Elisei F, Latterini L, Aloisi GG et al (2002) Excited-state properties and in vitro phototoxicity studies of three phenothiazine derivatives. Photochem Photobiol 75:11–21. doi:10.1562/0031-8655(2002) 0750011ESPAIV2.0.CO2 Escher B, Baumgartner R, Lienert J, Fenner K (2009) Predicting the ecotoxicological effects of transformation products. In: Boxall A (ed) Transformation products of synthetic chemicals in the environment. Springer, Heidelberg, pp 205–244 European Commission (2003) Technical Guidance Document on Risk Assessment Part II. Franke U, Munk A, Wiese M (1999) Ionization constants and distribution coefficients of phenothiazines and calcium channel antagonists determined by a pH-metric method and correlation with calculated partition coefficients. J Pharm Sci 88:89–95. doi:10.1021/js980206m Gocke E (1996) Review of the genotoxic properties of chlorpromazine and related phenothiazines. Mutat Res 366:9–21. doi:10.1016/ S0165-1110(96)90004-4 Harasko-van der Meer C, Bruecke T, Wenger S et al (1993) Two cases of long term dopamine D2 receptor blockade after depot neuroleptics. J Neural Transm 94:217–221. doi:10.1007/BF01277026 Heyes WF, Robinson ML (1985) HPLC procedure for the determination of some potential fluphenazine metabolites in urine. J Pharm Biomed Anal 3:477–481. doi:16/0731-7085(85)80063-7 Howard PH, Muir DCG (2010) Identifying new persistent and bioaccumulative organics among chemicals in commerce. Environ Sci Technol 44:2277–2285. doi:10.1021/es903383a Howard PH, Muir DCG (2011) Identifying new persistent and bioaccumulative organics among chemicals in commerce II: pharmaceuticals. Environ Sci Technol 45:6938–6946. doi:10.1021/es201196x HSDB (2011) Hazardous Substances Data Bank. US Department of Health and Human Services, National Institutes of Health, US National Library of Medicine: Bethesda, MD; http://toxnet.nlm. nih.gov/cgi-bin/sis/htmlgen?HSDB; http://toxnet.nlm.nih.gov/ cgi-bin/sis/htmlgen?HSDB. Accessed 23 Feb 2012 Huber C, Krause C (2006) Instruction Manual Fibox 3 Software Version 5.32. PreSens GmbH Regensburg Hubert-Roux M, Bounoure F, Skiba M et al (2010) Fragmentation pathways of metopimazine and its metabolite using ESI-MSn, HR-MS and H/D exchange. J Mass Spectrom 45:1121–1129. doi:10.1002/jms.1790 Jackson CJ, Hubbard JW, McKay G et al (1991) Identification of phase-I and phase-II metabolites of fluphenazine in rat bile. Intact glucuronide and sulfate conjugates. Drug Metab Dispos 19:188–193 Jeschke P (2004) The unique role of fluorine in the design of active ingredients for modern crop protection. Chembiochem A Eur J Chem Biol 5:570–589. doi:10.1002/cbic.200300833 Joyce C, Smyth W, Ramachandran V et al (2004) The characterisation of selected drugs with amine-containing side chains using electrospray ionisation and ion trap mass spectrometry and their determination by HPLC-ESI-MS. J Pharm Biomed Anal 36:465–476. doi:10.1016/j.jpba.2004.07.026 Key B, Howell R, Criddle C (1997) Fluorinated organics in the biosphere. Environ Sci Technol 31:2445–2454. doi:10.1021/es961007c Klopman G, Ivanov J, Saiakhov R, Chakravarti S (2005) MC4PC—an artificial intelligence approach to the discovery of structure toxic

Environ Sci Pollut Res (2012) 19:3162–3177 activity relationships (STAR). In: Helma C (ed) Predictive toxicology. Dekker, New York, pp 423–457 Kochevar IE, Hom J (1983) Photoproducts of chlorpromazine which cause red blood cell lysis. Photochem Photobiol 37:163–168. doi:10.1111/j.1751-1097.1983.tb04452.x Kolpin DW, Furlong ET, Meyer MT et al (2002) Pharmaceuticals, hormones, and other organic wastewater contaminants in U.S. streams, 1999–2000: a national reconnaissance. Environ Sci Technol 36:1202–1211 la Farre M, Perez S, Kantiani L, Barcelo D (2008) Fate and toxicity of emerging pollutants, their metabolites and transformation products in the aquatic environment. Trac-Trends Anal Chem 27:991– 1007. doi:10.1016/j.trac.2008.09.010 Lam M, Young C, Mabury S (2005) Aqueous photochemical reaction kinetics and transformations of fluoxetine. Environ Sci Technol 39:513–522. doi:10.1021/es0494757 Lindstrom AB, Strynar MJ, Libelo EL (2011) Polyfluorinated compounds: past, present, and future. Environ Sci Technol 45:7954– 7961. doi:10.1021/es2011622 López-Muñoz F, Alamo C, Cuenca E et al (2005) History of the discovery and clinical introduction of chlorpromazine. Ann Clin Psychiatry 17:113–135. doi:10.1080/10401230591002002 Meyer JS, Quenzer LF (2004) Psychopharmacology: drugs, the brain and behavior, 1st edn. Sinauer, New York Miolo G, Levorato L, Gallocchio F et al (2006) In vitro phototoxicity of phenothiazines: involvement of stable UVA photolysis products formed in aqueous medium. Chem Res Toxicol 19:156–163. doi:10.1021/tx0502239 Motten AG, Buettner GR, Chignell CF (1985) Spectroscopic studies of cutaneous photosensitizing agents–VIII. A spintrapping study of light induced free radicals from chlorpromazine and promazine. Photochem Photobiol 42:9–15. doi:10.1111/ j.1751-1097.1985.tb03540.x Mueller F (2000) Agrochemicals: composition, production, toxicology, applications, 1st edn. New York, Wiley OECD 301D (1992) Test no. 301D: ready biodegradability: Closed Bottle test. ISBN 9789264070349. OECD guidelines for the testing of chemicals, section 3: degradation and accumulation Prozialeck W, Cimino M, Weiss B (1981) Photoaffinity labeling of calmodulin by phenothiazine antipsychotics. Mol Pharmacol 19:264–269 Sackett PH, McCreery RL (1979) Effect of structure on phenothiazine cation radical reactions in aqueous buffers. J Med Chem 22:1447– 1453. doi:10.1021/jm00198a004 Schwabe U, Paffrath D (2010) Arzneiverordnungs-Report 2010: Aktuelle Daten, Kosten, Trends und Kommentare, 1st ed. Springer, Berlin Smyth WF (2003) Electrospray ionisation mass spectrometric behaviour of selected drugs and their metabolites. Anal Chim Acta 492:1–16. doi:10.1016/S0003-2670(02)01474-5 Smyth W (2005) Recent studies on the electrospray ionisation mass spectrometric behaviour of selected nitrogen-containing drug molecules and its application to drug analysis using liquid chromatography-electrospray ionisation mass spectrometry. J Chromatogr B 824:1–20. doi:10.1016/j.jchromb.2005.07.018 Trautwein C, Kümmerer K (2011) Incomplete aerobic degradation of the antidiabetic drug Metformin and identification of the bacterial dead-end transformation product Guanylurea. Chemosphere 85:765–773. doi:10.1016/j.chemosphere.2011.06.057 Trautwein C, Kümmerer K (2012) Degradation of the tricyclic antipsychotic drug chlorpromazine under environmental conditions, identification of its main aquatic biotic and abiotic transformation products by LC–MSn and their effects on environmental bacteria. J Chromatogr B 889–890:24–38. doi:10.1016/j.jchromb.2012.01.022 Tressaud A (2006) Fluorine and the environment: agrochemicals, archaeology, green chemistry & water, volume 2, 1st edn. Elsevier, New York

Environ Sci Pollut Res (2012) 19:3162–3177

US EPA (1999) United States Environmental Protection Agency, Federal Register: 40 CFR Part 372, Persistent bioaccumulative toxic (PBT) chemicals; final rule. 64:58665–58753 US EPA (2004) United States Environmental Protection Agency/ OPPT/Exposure Assessment Tools and Models/Estimation Program Interface (EPI) Suite Version 3.12 (August 17, 2004). http://www.epa.gov/oppt/exposure/pubs/episuite.htm. Accessed 13 Dec 2011 Viola G, Dall’Acqua F (2006) Photosensitization of biomolecules by phenothiazine derivatives. Curr Drug Targets 7:1135–1154. doi:10.2174/138945006778226561

3177 Viola G, Latterini L, Vedaldi D et al (2003) Photosensitization of DNA strand breaks by three phenothiazine derivatives RID G-36042010. Chem Res Toxicol 16:644–651. doi:10.1021/tx025680t Wesołowska O, Molnar J, Ocsovszki I, Michalak K (2009) Differential effect of phenothiazines on MRP1 and P-glycoprotein activity. In Vivo 23:943–947 WHO (2011) ATC/DDD Index, World Health Organisation Collaborating Centre for Drug Statistics Methodology. http:// www.whocc.no/atc_ddd_index. Accessed 15 Nov 2011 Wolfbeis OS (2002) Fiber-optic chemical sensors and biosensors. Anal Chem 74:2663–2678. doi:10.1021/ac020176e