Selective determination of antimycotic drugs in environmental water samples by mixed-mode solid-phase extraction and liquid chromatography quadrupole time-of-flight mass spectrometry

Journal of Chromatography A, 1339 (2014) 42–49

Contents lists available at ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Selective determination of antimycotic drugs in environmental water samples by mixed-mode solid-phase extraction and liquid chromatography quadrupole time-of-flight mass spectrometry J. Casado, I. Rodríguez ∗ , M. Ramil, R. Cela Departamento de Química Analítica, Nutrición y Bromatología, Instituto de Investigación y Análisis Alimentario (IIAA), Universidad de Santiago de Compostela, Santiago de Compostela 15782, Spain

a r t i c l e

i n f o

Article history: Received 28 January 2014 Received in revised form 24 February 2014 Accepted 25 February 2014 Available online 12 March 2014 Keywords: Antimycotic drugs Mixed-mode solid-phase extraction Water analysis Liquid chromatography quadrupole time-of-flight mass spectrometry

a b s t r a c t The suitability of mixed-mode (reversed-phase and cationic exchange) solid-phase extraction (SPE) for the selective concentration of basic antimycotic drugs (belonging to triazole, imidazole and allylamine chemical classes) in environmental water samples has been demonstrated for first time. The use of a sequential elution protocol, allowing the removal of neutral and acidic interferences before analytes extraction, led to a significant reduction of matrix effects, during electrospray ionization (ESI), in comparison with results reported for reversed-phase sorbents. In combination with liquid chromatography (LC) quadrupole time-of-flight (QTOF) mass spectrometry (MS) determination, the developed method attained limits of quantification (LOQs) comprised between 2 and 15 ng L−1 . After internal surrogate correction, accurate results (in most cases, recoveries ranged between 75 and 117%) were obtained for spiked aliquots of raw and treated wastewater, as well as river water, using quantification against calibration standard solutions in methanol (2% in NH3 ). Accurate, scan MS/MS spectra allowed the unambiguous identification of target compounds in environmental samples; furthermore, the information contained in MS spectra was used for the screening of additional antimycotics in the processed samples. Fluconazole, ketoconazole, miconazole and clotrimazole were measured in wastewater samples at concentrations up to 200 ng L−1 . The screening capabilities of the LC–QTOF-MS system permitted to identify the systematic presence of climbazole in the processed samples. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Antimycotics constitute a broad group of drugs designed to treat infections caused by fungi. Most of them are azole (triazoles and imidazoles) or allylamine compounds. In addition to topic applications, some antimycotics are administered orally and, less often, intravenous. Excretion of un-metabolized drugs added to wash off from treated skin areas result in the introduction of these compounds in municipal sewage water [1,2]. Given that antimycotics are designed to act as fungi killers, through inhibition of certain enzymes, they might also disturb the endocrine system of aquatic organisms. In fact, azolic compounds interfere with the production of aromatase enzymes, which are responsible for sexual differentiation of vertebrates during larval stages, and also for the balance between androgenic and estrogenic hormones in mammals [3].

∗ Corresponding author. Tel.: +34 881814387; fax: +34 881814468. E-mail address: [email protected] (I. Rodríguez). http://dx.doi.org/10.1016/j.chroma.2014.02.087 0021-9673/© 2014 Elsevier B.V. All rights reserved.

Among the most often prescribed antimycotic drugs, fluconazole (FCZ) has been detected at similar levels in the inlet and outlet streams of sewage treatment plants (STPs), which suggests resistance to biodegradation [4]. Less polar compounds, such as clotrimazole (CTZ), ketoconazole (KTZ), econazole (ECZ) and miconazole (MCZ), have been also found in wastewater entering municipal STPs [4–6]. Although their levels are significantly reduced in treated wastewater, sludge sorption, and not degradation, appears to be the main responsible for such reduction [7,8]. Analytical methods for the determination of antimycotic drugs in environmental water samples usually rely on liquid chromatography (LC) with tandem mass spectrometry (MS/MS), after solid-phase extraction (SPE) of water samples, using reversedphase type sorbents. With the exception of FCZ, most antimycotics display from moderate to high lipophilic properties; thus, there are prone to sorptive losses on filters and glassware material during SPE. Acidification of water samples results in analytes protonation, increasing their water solubility and minimizing sorption problems [7,9]. On the other hand, reducing the pH of water samples

0.5 0.5 0.7 0.7

0.5 1 1 0.9999 0.9998 0.9996

238.0783 204.9818 489.1455 242.1035 247.1449 69.0447 69.0447 93.0699 220.0681 158.9763 82.0530 165.0699 170.1018 125.0153 158.9763 141.0699 18 20 48 20 20 22 24 14 307.1113 328.0614 531.1560 277.0788 282.1092 383.0293 416.9904 292.2060

Precursor ion (Da)

Collision energy (eV)

Quantification ion (Da)

Other product ions (Da)

Linearity (R2 , 1–200 ng mL−1 , 8 levels)

increases the retention of acidic compounds on reversed-phase sorbents, resulting in more complex extracts than those obtained at neutral pH. The consequence of these too complex extracts is the existence of significant variations in the yield of electrospray ionization (ESI) for wastewater extracts versus calibration standards. In this sense, several studies have recognized the existence of large ionization suppression effects (up to 75%) after concentration of 100 mL of raw sewage water [5,9]. Although signal suppression can be compensated with isotopic labelled internal surrogates (IS), it results in increased limits of quantification (LOQs). Mixed-mode SPE sorbents permit increasing the selectivity of reversed-phase materials when applied to the concentration of ionizable compounds. Particularly, the OASIS MCX cartridges have provided cleaner extracts than OASIS HLB ones, when using a suitable elution protocol to fractionate basic analytes (e.g. drugs of abuse) from neutral and acidic interferences [10]. As far as we could trace, with regards to antimycotic drugs, the MCX cartridges have been tested only for the extraction of CTZ from wastewater. In this occasion, authors reported inappropriate results (the attained recoveries are not given); however, details regarding extraction and elution conditions are not provided [7]. The primary aim of this study was to assess the capability of the mixed-mode (reversed-phase and cationic exchanger) OASIS MCX sorbent for the extraction of basic antimycotic drugs from environmental water samples, improving the selectivity of the concentration process by means of a sequential elution strategy. As a secondary aim, we evaluated the use of a hybrid quadrupole time-of-flight (QTOF) MS instrument, instead of a triple quadrupole (QqQ) one, for the determination of target compounds after LC separation. The information contained in accurate, scan MS spectra provided by this system was used to screen the presence of additional antimycotics in the processed samples.

0.9994 0.9985 0.9980 0.9999

43 LOQsa (ng mL−1 )

J. Casado et al. / J. Chromatogr. A 1339 (2014) 42–49

10.7 12.7 13.0 13.5 13.6 14.1 15.1 15.3 0.4 3.6 4.3 4.1 4.1 5.5 6.1 5.6 a

Instrumental LOQs, without considering the SPE step.

11.01, 2.64 2.94 6.88 6.12 6.12 6.68 6.64 7.1 FCZ ETZ KTZ CTZ CTZ-d5 ECZ MCZ TRB Fluconazole Etaconazole Ketoconazole Clotrimazole Clotrimazole-d5 Econazole Nitrate Miconazole Nitrate Terbinafine hydrochloride

pKa Abbreviation Analyte

ECZ nitrate salt (100%), etaconazole (ETZ, 96.7%), CTZ (100%), FCZ (98%), KTZ (98%), (±)-MCZ nitrate salt (100%) and terbinafine hydrochloride (TRB, 98%) were obtained from Sigma (Milwaukee, WI, USA). CTZ-d5 (98%), used as IS, was acquired from Toronto Research Chemicals (North York, ON, Canada). Relevant properties (pKa and log Kow values) for above compounds are given in Table 1; whereas, their chemical structures are provided as supplementary information, Fig. S1. Individual solutions of each compound and the IS were dissolved in methanol. Further dilutions and mixtures of them were prepared in the same solvent. Calibration standard solutions were dissolved in methanol containing a 2% of NH3 . Methanol and acetonitrile, HPLC-grade purity; hydrochloric acid (37%), ammonia (25% solution in methanol) and ammonium acetate (99%) were supplied by Merck (Darmstadt, Germany). Ultrapure water was obtained from a Milli-Q (Millipore, Billerica, MA, USA) system. SPE cartridges (OASIS HLB 200 mg and OASIS MCX 150 mg) were acquired from Waters (Milford, MA, USA). Grab samples of raw and treated wastewater were obtained, in different dates, from the same STP, serving a population of 100.000 inhabitants in Galicia (Northwest Spain). The STP was equipped with primary and biological treatment units. Surface water was collected from the river receiving the effluent of this STP. Treated wastewater samples were also obtained from other STPs, in the same geographic area. Samples were taken in glass vessels. Immediately after reception, they were adjusted at pH 3, spiked with the IS and filtered before SPE concentration.

Table 1 Relevant properties and LC–QTOF-MS determination conditions for antimycotic drugs.

2.1. Standards, solvents, reagents and samples

Log Kow

Retention time (min)

2. Experimental

44

J. Casado et al. / J. Chromatogr. A 1339 (2014) 42–49

2.2. Sample preparation Extraction conditions were optimized with spiked aliquots of ultrapure, river and sewage water. Unless otherwise is stated, during optimization of SPE conditions an addition level of 1 ng mL−1 was used for all compounds. Breakthrough experiments were carried out passing the samples through two cartridges connected in series. Optimization of elution solvent volume was performed collecting consecutive fractions (2 mL each) from the SPE cartridge. Cartridges were conditioned with 5 mL of methanol followed by the same volume of ultrapure water, previously adjusted to the same pH as water samples. Before elution, cartridges were dried for 30 min with a gentle stream of nitrogen. Under optimized conditions, samples were adjusted at pH 3 and mixed with a 5% (v/v) of methanol. Thereafter, the considered volume (from 150 to 500 mL, depending on the matrix) was passed through a glass fibre filter. The filter was washed with 10 mL of methanol, which were combined with the filtrated solution. Analytes were concentrated with OASIS MCX cartridges at a flow of c.a. 5 mL min−1 . Thereafter, the sorbent bed was washed with 5 mL of a methanol:water (10:90) solution in order to release weakly retained polar interferences. After the drying step, neutral and acidic compounds were removed by rinsing the sorbent with 2.5 mL of methanol (0.1% in formic acid). Finally, antimycotic drugs were recovered with 2 mL of methanol containing a 2% (v/v) of NH3 . Some additional SPE extractions were performed using OASIS HLB cartridges with samples adjusted also at pH 3. Conditioning conditions were the same as those described for the MCX sorbent. The washing step was accomplished with 5 mL of a methanol:water solution (10:90) and, after drying the sorbent, analytes were eluted using 5 mL of methanol, which were further concentrated to 2 mL. 2.3. Determination conditions Determinations were carried out with a LC–ESI-QTOF-MS system acquired from Agilent (Wilmington, DE, USA). The LC instrument was an Agilent 1200 Series, consisting of an autosampler, two isocratic high pressure mixing pumps, a vacuum degasser unit and a chromatographic oven. The QTOF mass spectrometer was an Agilent 6520 model, furnished with a Dual-Spray ESI source. Compounds were separated in a Zorbax Eclipse XDB C18 column (100 mm × 2 mm, 3.5 ␮m) acquired from Agilent and connected to a C18 (4 mm × 2 mm) guard cartridge from Phenomenex (Torrance, CA, USA). Ultrapure water (A) and methanol (B), both 5 mM in ammonium acetate, were used as mobile phases applying the following gradient: 0–3 min, 5% B; 5 min, 93% B; 20–21 min, 100% B; 22–30 min, 5% B. The mobile phase flow was 0.2 mL min−1 , the injection volume for standards and sample extracts was 10 ␮L and the column temperature was set at 30 ◦ C. Nitrogen (99.999%), provided by a high purity generator (ErreDue srl, Livorno, Italy), was used as nebulizing (45 psi) and drying gas (350 ◦ C, 11 L min−1 ) in the ESI source. The QTOF instrument worked in the 2 GHz Extended Dynamic Range resolution mode (mass resolution 5000 at m/z values of 120). Analytes were quantified in ESI(+), applying a capillary voltage of 3000 V. A mass reference solution (Agilent calibration solution A) was continuously infused in the source of the QTOF system, through the second nebulizer, to guarantee the accuracy of m/z assignations. Recalibration of the mass axis was continuously performed considering ions with m/z values of 121.0509 and 922.0098. The Mass Hunter Workstation software was used to control the LC–ESI-QTOF-MS system and to process the obtained data. Precursor ions for target compounds were obtained using a fragmentor voltage of 150 V. Collision energies were optimized with the aim of generating, when possible, several products from each

precursor. Ion product scan (MS/MS) spectra were acquired in the range of m/z values from 55 to 550 units, considering a window of 1.5 min around the retention time of each analyte. MS scan spectra (m/z range from 100 to 1400 units) were simultaneously recorded to the MS/MS ones. Acquisition rates in MS and MS/MS modes were set at 2.5 spectra s−1 . Selective LC–MS and LC–MS/MS chromatograms were extracted with a mass window of 20 ppm, centred in the precursor and the most intense product ion of each antimycotic drug, respectively. The MS/MS mode was employed for quantification purposes. 2.4. Matrix effects, SPE efficiency and samples quantification Matrix effects (ME) were evaluated as follows: ME = [(Ase − Abe )/As ] × 100, where Ase is the response (peak area without IS correction) measured for a target compound in the spiked SPE extract from a water sample aliquot, Abe is the response for the same compound in an un-spiked extract of the same sample and finally, As is the response for a standard solution containing the same concentration of each analyte [11]. Thus, a ME value of 100% indicates the absence of changes between ionization yields for standard solutions and environmental water samples extracts. The efficiency of the SPE was calculated as the ratio between the responses (peak areas without IS correction) measured for spiked water samples (addition was made after pH adjustment and before filtration) and the SPE extracts from each sample, fortified after cartridge elution, multiplied by 100. The overall recoveries (R) of the procedure were defined as: R = [(Cs − Cb )/Ct ] × 100. Being Cs the concentration measured in the extract from a spiked water sample, Cb is the concentration in the extract from a non-spiked aliquot of the same sample and Ct is the concentration added to the sample. Cs and Cb were determined against calibration curves obtained for standard solutions of antimycotic drugs, prepared in methanol (2% in NH3 ), containing 75 ng mL−1 of CTZ-d5 . The levels of target analytes in environmental samples were quantified using these calibration standard solutions. 3. Results and discussion 3.1. LC–ESI-QTOF-MS determination parameters LC and ESI variables were optimized with the aims of improving the separation of antimycotic drugs and to maximize their response in the MS mode, respectively. All compounds render the typical cluster of signals for the [M+H]+ ion, with the known exception of CTZ, which lost the imidazole moiety in the ESI source; thus, the [M−C3 H3 N2 ]+ species was selected as its precursor ion. In case of ETZ, two peaks with the same spectrum and relative intensities 5:95 were obtained. Only, the most intense one was used for quantification purposes. Analytes retention in the LC column was affected not only by the mobile phase gradient, but also by the modifier. The retention times of TRB and the four imidazole drugs shifted to lower values using formic acid (0.1%), instead of ammonium acetate (5 mM), as modifier; whereas, the retention of FCZ and ETZ remained unaffected by the pH of the mobile phase. Ammonium acetate provided a better chromatographic separation and thus was adopted as modifier. Collision energies were optimized with the aim of obtaining several intense fragments from each precursor. The most intense product ions for each compound are provided in Table 1. The linearity in the response of the system was evaluated with standards in methanol, at eight different concentration levels, in the range from 1 to 200 ng mL−1 , maintaining the IS at 75 ng mL−1 . The plots of corrected peak areas versus concentration followed a linear model, with determination coefficients (R2 )

J. Casado et al. / J. Chromatogr. A 1339 (2014) 42–49

varying between 0.9980 and 0.9999, Table 1. The instrumental LOQs, defined as the concentration of each compound rendering a peak area 10 times higher than the standard deviation of the baseline noise in the MS/MS mode, ranged from 0.5 to 1 ng mL−1 . 3.2. Optimization of SPE conditions As reported for CTZ [7], lipophilic antimycotic drugs are prone to sorption on glass material and, particularly, in filters, which might led to reduced recoveries for SPE extractions and, also, to the report of false negatives when analyzing real samples. Since these compounds are positively charged at acidic pHs, the effect of sample pH in the efficiency of the SPE step, and in the potential losses occurring during filtration of environmental samples, was assessed. Comparison between recoveries provided by the MCX sorbent for ultrapure water samples (500 mL) at pH 6 and 3 revealed similar average values, without noticeable breakthrough problems for any compound (see supplementary data, Table S1). Thus, the cationic exchange mechanism is not required for the effective retention of target compounds in the mixed-mode sorbent. Standard deviations (SDs) of recoveries were relatively high in both cases, and particularly at neutral pHs (Table S1), likely due a higher significance of the above reported sorption problems. Thus, in further experiments, methanol (5%, v/v) was added to water samples to increase the solubility of the compounds. Losses during filtration were assessed by adding the compounds to 150 mL aliquots of a raw wastewater, containing a 5% of methanol. After a 10 min stirring step, samples were filtered and submitted to SPE extraction with MCX cartridges. Filters were soaked with 10 mL of methanol and the obtained extracts concentrated to 2 mL. Table S2 shows the relative response for each compound in the filter extract versus its total peak area. With the exception of the most polar triazolic compounds (FCZ and ETZ), the rest of analytes were lost in a significant extend during filtration of non-acidified (pH 7.8) wastewater aliquots. Losses were reduced to a maximum of 15% for MCZ when compounds were spiked over aliquots of the same wastewater sample, previously adjusted at pH 3. Analytes retention in glass wool filters was further reduced by rinsing the filters with 10 mL of methanol, which were mixed with the water sample before SPE, Table S2. In further experiments, water samples, containing a 5% of methanol, were adjusted at pH 3 before filtration. Thereafter, the glass wool filter was rinsed with 10 mL of methanol, which were mixed with the filtrated sample before SPE concentration. A significant advantage of mixed-mode sorbents is the possibility to fractionate neutral and ionizable species. In this work, after the loading step, cartridges were first rinsed with 5 mL of methanol:water (10:90) to remove poorly retained species. After drying the sorbent, acidified (0.1% in formic acid) methanol (2.5 mL) was passed through the cartridge in order to remove neutral interferences, at the same time that basic drugs are fixed through electrostatic interactions. Thereafter, consecutive fractions (2 mL each) of methanol, containing different percentages of NH3 , were considered for elution of the analytes. Fractionated elution experiments were carried out with ultrapure and raw wastewater, with similar results in both cases. The percentage of NH3 added to methanol affected the elution profile of antimycotic drugs. Considering a 0.5% of NH3 , two fractions of methanol (4 mL) were required for analytes elution. When the percentage of NH3 was increased to 2%, compounds were recovered with only 2 mL of solvent. The recoveries of the optimized SPE procedure are compiled in Table 2. With the only exception of ETZ, obtained values stayed above 80% considering sample volumes comprised between 150 mL (raw wastewater) and 500 mL (river water). ME were evaluated for the method optimized in this study and also considering SPE with

45

HLB cartridges, which were conditioned and eluted as reported in the experimental section. Table 3 summarized the ME data corresponding to different water samples passed through the MCX cartridges under optimized conditions. The last two columns in Table 3 contains ME values for raw wastewater using HLB cartridges, and MCX ones, which were eluted directly with 2 mL of methanol (2% in NH3 ), without considering the rinsing step with acidified methanol. Under optimized conditions, for most of the compounds, the ME varied between 87% and 112% in the three investigated samples, pointing out to the existence of just small changes in the yield of ESI ionization among the extracts from real samples and standards prepared in methanol (2% in NH3 ), Table 3. In case of FCZ, a signal attenuation around 45% was observed in case of wastewater samples. ME obtained for aliquots of the same raw wastewater using OASIS HLB cartridges and MCX ones, but without the fractionated elution protocol, were similar, ranging from 25 to 81%, Table 3. These signal attenuation levels match with those previously published for reversed-phase sorbents [5]. ME for river and treated wastewater samples concentrated with the reversed-phase OASIS HLB sorbent, are provided as supplementary information (Table S3). Even for these matrices, higher signal attenuation values were noticed after OASIS HLB concentration than using the MCX sorbent. 3.3. Performance of the method The overall recoveries of the reported method, after IS correction, were calculated for different environmental samples, considering two addition levels for each matrix. Recoveries obtained for TRB and imidazole drugs ranged between 75 and 111%, with associated SDs below 12%, except in the case of river water spiked at the lowest concentration level (Table 4). Globally, recoveries compiled in Table 4 for imidazole drugs are better than those published for the same compounds (from 34 to 116%) using reversed-phase HLB cartridges, considering the same IS as in this study [5]. For the more polar triazolic drug FCZ, the corrected recoveries displayed a larger variability, with a minimum average recovery of 63% in one of the samples. Procedural blanks, corresponding to the concentration of 500 mL of ultrapure water, displayed only a small peak at the retention time of MCZ (Fig. 1). Thus, for most compounds, the achieved LOQs were controlled by the sensitivity of the LC–QTOF-MS system and the enrichment factor provided by the sample preparation process. In the case of MCZ, the LOQ was calculated as 10 times the SD of its response in five consecutive procedural blanks, divided by the slope of its calibration curve and the concentration factor for each water matrix. For the rest of compounds, the procedural LOQs were estimated as 10 times the SDs of their chromatographic baseline noise divided by the slopes of their calibration curves and the concentration factor corresponding to each water sample matrix. Obtained LOQs ranged from 2 to 15 ng L−1 (Table 4). These values are in the same order of magnitude as the LOQs reported by Lindberg et al. [9] after concentration of 500 mL samples followed by LC–ESI-MS/MS using an ion-trap mass spectrometer (LOQs from 5 to 100 ng L−1 ). They are also equivalent to LOQs reported for on-line SPE combined with LC–ESI-MS/MS in a triple quadrupole instrument [4], and off-line SPE followed by the same determination technique (LOQs from 2 to 6 ng mL−1 for wastewater) [5]. 3.4. Real samples analysis The developed method was applied to six pairs of grab wastewater samples, simultaneously collected in the inlet and outlet streams of the same STPs, during different weeks in July and November 2013. Also, two samples of river water were obtained 3 Km downstream the discharge of the STP in November 2013.

46

J. Casado et al. / J. Chromatogr. A 1339 (2014) 42–49

Fig. 1. LC–ESI-MS/MS extracted chromatograms and spectra for a procedural blank, dotted line, and compounds detected in a raw wastewater sample (code 12, Table 5), solid line.

J. Casado et al. / J. Chromatogr. A 1339 (2014) 42–49

47

Table 2 Recoveries (%) of the optimized SPE process using mixed-mode cartridges, n = 3 replicates. Average recovery (%) ± SD

Compound

River water, 500 mL FCZ ETZ KTZ CTZ ECZ MCZ TRB

97 88 84 96 94 104 93

± ± ± ± ± ± ±

Treated wastewater, 300 mL

3 2 6 1 2 5 1

109 71 84 85 88 81 84

± ± ± ± ± ± ±

Raw wastewater, 150 mL

12 3 7 6 2 8 5

90 72 92 80 89 82 85

± ± ± ± ± ± ±

16 8 7 6 9 1 7

Table 3 Matrix effects evaluation for different water samples using the mixed-mode MCX sorbent under optimized conditions, n = 3 replicates. Matrix effect (%) ± SD

Compound

River water FCZ ETZ KTZ CTZ ECZ MCZ TRB a b

99 90 109 107 106 107 112

± ± ± ± ± ± ±

Treated wastewater

4 15 6 16 5 4 5

53 104 106 109 110 103 110

± ± ± ± ± ± ±

Raw wastewatera

Raw wastewater

2 1 3 3 11 5 4

58 97 93 99 101 101 87

± ± ± ± ± ± ±

2 3 2 3 3 5 7

25 52 38 54 39 60 81

± ± ± ± ± ± ±

Raw wastewaterb

3 2 3 2 4 4 6

27 60 47 58 49 66 75

± ± ± ± ± ± ±

1 2 2 2 4 2 1

Data obtained for reversed-phase OASIS HLB cartridges. Data obtained for mixed-mode OASIS MCX cartridges without sequential elution.

Table 4 Overall recoveries of the optimized method, after IS correction, and procedural LOQs (ng L−1 ) for different water matrices. Compound

River water, 500 mL a

FCZ ETZ KTZ CTZ ECZ MCZ TRB a

0.2 ng mL−1

104 94 90 102 100 111 99

± ± ± ± ± ± ±

5 2 10 4 5 6 3

a

Treated wastewater, 300 mL

0.05 ng mL−1

113 103 105 95 84 85 94

± ± ± ± ± ± ±

18 13 18 11 6 20 7

LOQs (ng L−1 )

a

0.5 ng mL−1

2 2 2 2 2 5 4

63 90 80 96 92 103 94

± ± ± ± ± ± ±

a

4 2 2 2 3 4 2

0.1 ng mL−1

117 101 100 101 105 96 101

± ± ± ± ± ± ±

12 1 8 5 1 7 4

Raw wastewater, 150 mL LOQs (ng L−1 ) 3 3 5 3 3 10 7

a

0.5 ng mL−1

86 82 75 104 90 92 101

± ± ± ± ± ± ±

a

9 5 12 2 1 5 3

0.1 ng mL−1

80 76 75 88 85 81 97

± ± ± ± ± ± ±

10 8 7 4 5 5 6

LOQs (ng L−1 ) 7 7 9 7 7 15 13

Concentrations added to water samples.

Furthermore, the effluents from three different STPs were taken in July 2013. ECZ, TRB and ETZ remained below the LOQs of the method in all samples. Concentrations measured for the rest of species are compiled in Table 5. FCZ was the compound displaying the highest concentrations in treated and river water samples, pointing out to

a poor biodegradability and relatively high mobility in the aquatic environment. In some pairs of samples, the concentrations of this compound in treated wastewater samples are higher than in the influent of the plant. This could be explained since data are referred to non-integrated samples. Another possibility is that, FCZ, which

Table 5 Concentrations (ng L−1 ) of antimycotic drugs in environmental water samples, n = 3 replicates. Code

Type

Sampling date

Conc. (ng L−1 ) ± SD

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

R.W. T.W. R.W. T.W. R.W. T.W. R.W. T.W. R.W. T.W. River R.W. T.W. River T.W. T.W. T.W.

05/07/13 05/07/13 12/07/13 12/07/13 04/11/13 04/11/13 12/11/13 12/11/13 21/11/13 21/11/13 21/11/13 27/11/13 27/11/13 27/11/13 12/07/13 12/07/13 12/07/13

86 72 83 95 20 40 93 37 41 70 32 59 69 25 28 31 16

FCZ

R.W., raw wastewater; T.W., treated wastewater.

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

5 12 3 9 4 14 17 2 10 7 4 22 9 8 8 7 4

KTZ

CTZ

72 ± 7 23 ± 7 110 ± 10 15 ± 3 102 ± 23 34 ± 17 191 ± 15 36 ± 7 55 ± 6 10 ± 3 4±1 58 ± 11 17 ± 1 11 ± 1
47 5 11 6 48 5 49 11 80 9 7 74 8 9 3 15 8

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

MCZ 6 1 1 1 10 1 7 1 3 1 1 15 1 2 1 2 1

30 ± 10
48

J. Casado et al. / J. Chromatogr. A 1339 (2014) 42–49

Fig. 2. (A) LC–ESI-MS extracted chromatogram and ESI(+) MS spectrum for a peak tentatively identified as climbazole. (B) MS/MS spectrum for the same peak.

can be orally administered, arrives to the STP as free and conjugated compound, undergoing de-conjugation at the STP. Existence of relatively high levels of FCZ residues in treated wastewater is in agreement with the results reported in the literature [4,5,9]. FCZ was also detected in the effluents from other STPs (codes 15–17, Table 5) at higher levels than the rest of antimycotics. KTZ, CTZ and MCZ were also ubiquitous in raw wastewater with maximum concentrations close to 200 ng L−1 for the first species (sample code 7, Table 5). Their levels in treated samples were significantly lower, particularly in case of MCZ, remaining below the LOQs of the method in several cases, Table 5. Fig. 1 shows the extracted LC–MS/MS chromatograms for a procedural blank and sample code 12, Table 5. The accurate ion product scan MS/MS spectrum of each peak provided an unambiguous confirmation of its identity. Analysis of two river water samples (codes 11 and 14), confirmed the presence of detectable concentrations of the above four compounds, with the highest concentrations corresponding again to FCZ. 3.5. Screening for additional antimycotic drugs LC–QTOF-MS chromatograms were screened for other antimycotic drugs, not considered during method development, but included in the formulation of pharmaceuticals distributed by the public medical service in Spain, and/or in personal care products. The list of considered species is provided as supplementary information, Table S4. The procedure used in this post-target study

has been described in previous articles [12,13]. In brief, LC–MS chromatograms were explored for the [M+H]+ ions of compounds compiled in Table S4, within a mass window of 20 ppm. In case of a positive finding, the MS/MS scan spectrum for the candidate peak was acquired, using different collision energies. Climbazole was the only compound detected in the extracts from environmental water samples. Fig. 2A shows the LC–MS chromatogram for the characteristic [M+H]+ ion of climbazole (293.1051 Da), corresponding to the extract from a treated wastewater sample (code 13, Table 5) and the MS spectrum for this peak. The boxes in red (Fig. 2A) represent the theoretical ESI(+) spectrum of climbazole. The overall score of its MS spectrum was 92 over 100, and the mass error remained below 1 ppm. The scan MS/MS spectrum for this peak is shown in Fig. 2B. The nominal masses of the two most intense product ions (69 and 197 Da) matched with those used for the quantification of climbazole in MRM methods, developed for LC-QqQ instruments [14]. Furthermore, the MS/MS fragmentation pattern depicted in Fig. 2B (it reflected losses of the positively charged imidazole ring and the carbonyl moiety) is equivalent to that published by our group for triadimefon [15]. Climbazole and the agricultural pesticide triadimefon share the same chemical structure, with the only difference that the triazolic ring of the latter is replaced by an imidazole moiety in the former. Finally, the identity of climbazole was confirmed by injection of a commercial standard. Climbazole is not employed as active component in prescription drugs, but as ingredient of personal care products, particularly in shampoos. Recent articles have demonstrated that this fungicide is phytotoxic [16] and persistent in sludge-amended soils [17]. These

J. Casado et al. / J. Chromatogr. A 1339 (2014) 42–49

data, added to detection of climbazole in all samples compiled in Table 5, indicate that it should be considered in further quantitative methods. 4. Conclusions Mixed-mode SPE of environmental water samples, combined with a sequential elution protocol, improved the selectivity of antimycotic drugs extraction from water samples, attaining a relevant reduction of ESI signal attenuation effects when compared to conventional reversed-phase extraction. Acidification of samples and addition of methanol is mandatory to avoid losses of the analytes during filtration of sewage water. The developed mixed-mode SPE protocol followed by LC–QTOF-MS determination provided LOQs low enough for the selective and unambiguous determination of target compounds in STPs. Also, the information derived from accurate, MS scan spectra, recorded through LC chromatograms, can be used in post-target screening studies with the aim of detecting the presence of additional antimycotics in the aquatic environment, as demonstrated in this work for climbazole. Data obtained for real samples confirmed the presence of MCZ, CTZ, KTZ and FCZ in raw wastewater samples, with the latter species measured also at similar levels in treated wastewater. Acknowledgements This study has been supported by the Spanish Government and E.U. FEDER funds (project CTQ2012-33080). We thank to Viaqua for providing access to sewage water samples.

49

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.chroma.2014.02.087. References [1] O. Sui, B. Wang, W. Zhao, J. Huang, G. Yu, S. Deng, Z. Qiu, S. Lu, Chemosphere 89 (2012) 280. [2] F. Heeb, H. Singer, B. Pernet-Coudier, W. Qi, H. Liu, P. Longrée, B. Müller, M. Berg, Environ. Sci. Technol. 46 (2012) 8680. [3] J.A. Zarn, B.J. Bruschweiler, J.R. Schalatter, Environ. Health Perspect. 111 (2003) 255. [4] M. Kahle, I.J. Buerge, A. Hauser, M.D. Müller, T. Poiger, Environ. Sci. Technol. 42 (2008) 7193. [5] Q. Huang, Y. Yu, C. Tang, X. Peng, J. Chromatogr. A 1217 (2010) 3481. [6] X. Peng, Q. Huang, K. Zhang, Y. Yu, Z. Wang, C. Wang, Sci. Total Environ. 426 (2012) 311. [7] M. Peschka, P.H. Roberts, T.P. Knepper, Anal. Bioanal. Chem. 389 (2007) 959. [8] A.I. García-Valcárcel, J.L. Tadeo, J. Sep. Sci. 34 (2011) 1228. [9] R.H. Lindberg, J. Fick, M. Tysklind, Water Res. 44 (2010) 649. ˜ J.B. Quintana, I. Rodríguez, M. González-Díez, R. Cela, Anal. [10] I. González-Marino, Chem. 84 (2012) 1708. [11] B.K. Matuszewski, M.L. Constanzer, C.M. Chavez-Eng, Anal. Chem. 75 (2003) 3019. ˜ J.B. Quintana, I. Rodríguez, N. Sánchez-Méndez, R. Cela, [12] I. González-Marino, Anal. Bioanal. Chem. 404 (2012) 3135. [13] R. Celano, I. Rodríguez, R. Cela, L. Rastrelli, A.L. Piccinelli, Talanta 118 (2014) 312. [14] G. Chen, M. Hoptroff, X. Frei, Y. Su, H.G. Janssen, J. Chromatogr. A 1317 (2013) 155. [15] T. Rodríguez-Cabo, I. Rodríguez, M. Ramil, R. Cela, Anal. Bioanal. Chem. 405 (2013) 4187. [16] E. Richter, A. Wick, T.A. Ternes, A. Coors, Environ. Toxicol. Chem. 32 (2013) 2816. [17] Z.F. Chen, G.G. Ying, Y.B. Ma, H.J. Lai, F. Cheng, C.G. Pan, J. Agric. Food Chem. 61 (2013) 6198.