Anal Bioanal Chem (2014) 406:5765–5774 DOI 10.1007/s00216-014-8026-y

PAPER IN FOREFRONT

Biosynthesis of 15N-labeled cylindrospermopsin and its application as internal standard in stable isotope dilution analysis Katrin Kittler & Holger Hoffmann & Franziska Lindemann & Matthias Koch & Sascha Rohn & Ronald Maul

Received: 26 May 2014 / Revised: 7 July 2014 / Accepted: 8 July 2014 / Published online: 27 July 2014 # Springer-Verlag Berlin Heidelberg 2014

Abstract Cylindrospermopsin (CYN) is a cyanobacterial toxin associated with human and animal poisonings. Due to its toxicity in combination with its widespread occurrence, the development of reliable methods for selective, sensitive detection and accurate quantification is mandatory. Liquid chromatography tandem mass spectrometry (LC-MS/MS) analysis using stable isotope dilution analysis (SIDA) represents an ideal tool for this purpose. U-[15N5]-CYN was synthesized by culturing Aphanizomenon flos-aquae in Na15NO3-containing cyanobacteria growth medium followed by a cleanup using graphitized carbon black columns and mass spectrometric characterization. Subsequently, a SIDA-LC-MS/MS method for the quantification of CYN in freshwater and Brassica matrices was developed showing satisfactory performance data. The recovery ranged between 98 and 103 %; the limit of quantification was 15 ng/L in freshwater and 50 μg/kg dry weight in Brassica samples. The novel SIDA was applied for CYN determination in real freshwater samples as well as in kale and in vegetable mustard exposed to toxin-containing irrigation water. Two of the freshwater samples taken from Electronic supplementary material The online version of this article (doi:10.1007/s00216-014-8026-y) contains supplementary material, which is available to authorized users. K. Kittler : H. Hoffmann : F. Lindemann : M. Koch : R. Maul BAM Federal Institute for Materials Research and Testing, Richard-Willstätter-Straße 11, 12489 Berlin, Germany S. Rohn Hamburg School of Food Science, Institute of Food Chemistry, University of Hamburg, Grindelallee 117, 20146 Hamburg, Germany R. Maul (*) Leibniz-Institute of Vegetable and Ornamental Crops Großbeeren/ Erfurt e.V., Theodor-Echtermeyer-Weg 1, 14979 Großbeeren, Germany e-mail: [email protected]

German lakes were found to be CYN-contaminated above limit of quantification (17.9 and 60.8 ng/L). CYN is systemically available to the examined vegetable species after exposure of the rootstock leading to CYN mass fractions in kale and vegetable mustard leaves of 15.0 μg/kg fresh weight and 23.9 μg/kg fresh weight, respectively. CYN measurements in both matrices are exemplary for the versatile applicability of the developed method in environmental analysis. Keywords Cyanotoxin . Quantification . Surface water . Vegetable plants . SIDA . HPLC-MS/MS

Introduction Toxic substances produced by cyanobacteria represent an emerging issue for food and feed safety as well as for environmental toxicology. One of these toxins is cylindrospermopsin (CYN) (Fig. 1), a cyanotoxin mainly causing damage to liver and other organs including kidney, lungs, and heart [1]. A large variety of cyanobacterial species of the genera Cylindrospermopsis, Aphanizomenon, and Anabaena is capable of producing CYN. CYN occurrence has been reported from almost all continents. Because CYN-forming species only populate freshwater or water bodies with low salinity, CYN predominantly occurs in lakes, rivers, ponds, and reservoirs [2]. The CYN concentrations that have been observed in water bodies varied strongly depending mainly on the climatic region. While in Australian ponds, concentrations higher than 500 μg/L have been reported, a maximum of 12.1 μg/L total CYN was detected in an investigation of 21 temperate German lakes [3, 4]. Potential human exposure routes are diverse: A direct CYN exposure may arise from accidental uptake of water when swimming in contaminated recreational water

5766

Fig. 1 Chemical structure of cylindrospermopsin

bodies or is possible by ingestion of improperly treated drinking water [1]. Indirect exposure to CYN has to be expected by ingestion of contaminated food of aquatic origin such as bivalves, fish, edible algae, or crop plants irrigated with contaminated water [3, 5–7]. A guideline safety value of 1 μg/L was proposed in 2003 for drinking water and was adopted in a few countries, e.g., New Zealand [8, 9]. Because of the widespread occurrence of CYN and its toxic potential, the development of sensitive and selective analytical methods is required. The spectrum of existing detection methods ranges from a nonspecific mouse bioassay [10], which was used for determination of the toxicity of biomass deriving from four Cylindrospermopsis strains in Portugal, a sensitive ELISA assay, but also LC-UV methods [11, 12]. Finally, liquid chromatography mass spectrometry (LC-MS) methods have been described for a selective and sensitive determination of CYN in different matrices [7, 13, 14]. One of the liquid chromatography tandem mass spectrometry (LC-MS/MS) methods available is designed for the determination of CYN in freshwaters and fish muscle showing satisfactory recoveries (limits of quantification (LOQs) of 0.1 and 1.0 ng/g, respectively) and repeatability; however, it did not use an internal standard (ISTD) or matrix-matched calibration [15]. The latter fact hampers a more versatile applicability of the described LC-MS/MS method for CYN quantification in different matrices. Nevertheless, the extraction of CYN even from pure water may cause some problems due to a considerable matrix background which is known to lead to ion suppression or enhancement [16–18]. Apart from matrix effects, sample treatment influences the quantifiable amount of CYN. A sample treatment including cleanup or pre-concentration steps is often required for complex matrices or water samples contaminated at low levels, respectively. Due to the high polarity and zwitterionic character of CYN, cleanup methods are limited, but solid-phase extraction (SPE) techniques with carbon-based sorbents have been applied successfully [14, 19–21]. However, a potential loss of the analyte needs to be considered as soon as cleanup steps are included.

K. Kittler et al.

Matrix effects and losses due to sample cleanup are typically compensated by using an ISTD. In the case of CYN, only one method has been published so far using HEPES, an ethanesulfonic acid, as ISTD [22]. This substance possesses chemical characteristics different from CYN; however, it was found to be a suitable ISTD for the analysis of CYN in water samples or cyanobacterial extracts. In order to establish LC-MS/MS as a versatile reference method for CYN analysis, a chemically more similar ISTD is desirable. An isotopically labeled ISTD is identical to the native analyte with regard to structure and chemical behavior and therefore appropriate to compensate losses of the analyte during sample preparation and any kind of matrix effect [17]. Recently, Sano et al. succeeded in biosynthesizing a fully 15N-labeled standard for some microcystins, which was subsequently applied in stable isotope dilution analysis (SIDA) [18]. The aim of this study was to develop a sensitive, selective, and accurate method to quantify CYN in representative matrices based on SIDA-LC-MS/MS. A prerequisite was the biosynthesis and characterization of fully 15N-labeled CYN. Biosynthesized U-[15N5]-CYN was applied in the analysis of vegetable plant and freshwater samples.

Materials and methods Materials A reference solution of CYN in water of 9.8 mg/L (initial concentration of the certified standard in water: (30±2 μM); National Research Council-Institute for Marine Biosciences, Halifax, Canada) was donated from the Federal Environmental Agency. CYN of noncertified purity was provided by Santa Cruz Biotech (Heidelberg, Germany). Na15NO3 was purchased from Cambridge Isotope Laboratories (Tewksbury, USA). Methanol (MeOH) and acetonitrile (ACN) were of highperformance liquid chromatography (HPLC) grade and obtained from Th. Geyer (Berlin, Germany). Ammonium acetate was purchased from J.T. Baker (Deventer, The Netherlands). K2HPO4 and ZnSO4 × 7 H2O were purchased from Carl Roth (Karlsruhe, Germany). Na14NO3, FeCl3 × 6 H2O, NaHCO3, MnSO4 × H2O, and trifluoroacetic acid (TFA) were obtained from Sigma Aldrich (Taufkirchen, Germany). Citric acid, MgSO4 × 7 H2O, CuSO4 × 5 H2O, CaCl2 × 2 H2O, Na2CO3, H3BO3, CuSO4 × 5 H2O, acetic acid, and formic acid were provided by Merck (Darmstadt, Germany). Graphitized carbon black (GCB) was purchased from Agilent SampliQ (Lake Forest, Canada). Deionized water was supplied by a Seralpur PRO 90CN system (RansbachBaumbach, Germany).

15

N-labeled cylindrospermopsin and its application

Methods Biosynthesis and preparation of U-[15N5]-CYN The nonaxenic cyanobacterial strain Aphanizomenon flos-aquae 22D11 (APH) isolated from the lake “Heiliger See,” Berlin, Germany [23] was cultured nonsterile in gas washing bottles under a 16+8-h light/ dark cycle. Growth medium consisted of macronutrients contained in Z medium for cyanobacteria with some alterations [24]. Instead of using Na14NO3, Na15NO3 was added to the medium; EDTA was replaced by 1.6 μM citric acid. The composition of micronutrients was adopted from Kuhl and Lorenzen [25]. The culture was constantly aerated with indoor air. Once a month, the medium was renewed and tested for U-[15N5]-CYN formation. The volume of the precultures was approximately 75 mL. After 18 months, the culture volume was scaled up to 500 mL for U-[15N5]-CYN isolation from the culture medium. The collected medium was separated from the cyanobacterial cells by centrifugation (1,495×g) and enriched by SPE using GCB (1 g) as sorbent material. After conditioning self-packed SPE cartridges using 20 mL MeOH/TFA (99/ 1; v/v) and 30 mL of water, 500 mL medium was loaded on a cartridge. The cartridge was washed by adding 20 mL of water and brought to dryness by applying a vacuum. U-[15N5]-CYN was eluted with 20 mL MeOH/TFA (99/1; v/v). After evaporation to dryness using a rotary evaporator Rotavapor R-200 (Büchi, Flawil, Switzerland), the residue was dissolved in 500 μL water for the next cleanup step. The cleanup was performed with a KNAUER HPLC system (Berlin, Germany) equipped with a quaternary gradient pump (K-1001), autosampler Midas (Spark, Emmen, The Netherlands), diode array detector (DAD) (K-2700), and automatic fraction collector Foxy R1 (Teledyne Isco, Lincoln, USA) using a 150×3-mm particle size 5 μm Luna phenylhexyl column with guard protection (Phenomenex Ltd., Aschaffenburg, Germany). Mobile phase A consisted of water and mobile phase B consisted of MeOH, both containing 0.5 % formic acid (v/v). An isocratic elution using 98 % A and 2 % B for CYN was applied. The column was flushed with 20 % A and 80 % B starting from minute 5.5 followed by an equilibration step for 17.5 min. The flow rate was 0.5 mL/min and the injection volume was 50 μL. DAD detection was set to λ=262 nm. The CYN-containing fractions (4.7–5.7 min) were evaporated to dryness using a rotary evaporator and reconstituted in 1 mL MeOH. Structure confirmation was done by LC hyphenated to tandem mass spectrometry (LC-MS/MS) comparing fragmentation pattern and retention time to a native CYN calibration standard and by high-resolution mass spectrometry (HRMS) determining the exact mass of labeled CYN. The amount of

5767

the biosynthesized and purified U-[15N5]-CYN was estimated by comparing the obtained LC-MS/MS signal intesity to a native CYN calibration standard. Freshwater sample preparation In summer 2013, water samples from different German and Kenyan freshwater sources were collected on the shoreline to determine CYN in environmental samples: “Rangsdorfer See” (Teltow-Fläming, Germany), “Motzener See” (DahmeSpreewald, Germany), “Seddiner See” (Potsdam-Mittelmark, Germany), “Peene-Strom” (Mecklenburg-Vorpommern, Germany), “Pond Juja” (Kenya), “Pond Juja” irrigation water (Kenya), and local water supply sample (Kenya). The collected water was stored in amber glass bottles at 4 °C in the dark until analysis. Samples (100 mL, n=3) were spiked gravimetrically with U-[15N5]-CYN ISTD (final concentration approximately 30.0 μg/L) before further sample preparation. Naturally CYN-free freshwater samples (100 mL, n=3) were spiked gravimetrically with CYN calibration standard (STD) and ISTD yielding CYN concentrations in nonconcentrated samples of 10.0, 1.0, and 0.03 μg/L and a final U-[15N5]-CYN ISTD concentration in concentrated samples of approximately 30.0 μg/L for method development. Environmental and spiked freshwater samples were equilibrated with the spiked standards for at least 0.5 h and then filtered using a 0.45-μm regenerated cellulose membrane followed by SPE (GCB, 1 g). SPE procedure was the same as described for enrichment of the labeled standard. Deviating from the procedure previously described, for evaporation, an RVC 2-25 CD plus vacuum concentrator (Christ, Osterode am Harz, Germany) was used. Concentration factors were 10, 100, or 1,000 depending on spiked native CYN concentration (see Table 1) and 1,000 for environmental freshwater samples. Vegetable plant sample preparation For the preparation of naturally CYN-contaminated vegetable plants, kale seeds (Brassica oleracea var. sabellica) and vegetable mustard seeds (Brassica juncea) were grown in a hydroponic system as described by Kittler et al. [7]. The plant treatment was conducted by irrigation of the plants with a nutrient solution containing a cyanobacterial extract. The CYN concentration in the nutrient solution contaminated with CYN was adjusted to 75 μg/L (verified by LC-MS analysis). Control plants were irrigated with nutrient solution without cyanobacterial extract. After 21 days, vegetable plants from each species and each treatment were harvested. The leaves were weighed and dependent on treatment randomly pooled (three plants = one sample). The subsequent sample preparation was conducted according to Kittler et al. [7]. For CYN determination, 40 mg of freeze-dried samples of kale and vegetable mustard (three samples of each treatment with

5768 Table 1 Identified CYN molecules with different numbers of stable 15N isotopes in the CYN standard isolated from 15N-containing cyanobacterial growth medium by LC-HRMS. Percentage of completely 14N CYN related to the different stages (x) of 15 N incorporation in the CYN molecule

K. Kittler et al.

Contained 15N atoms (x) in CYN

Sum formula [M + H]+

Determined mass (amu)

Error (ppm)

U −½14 N5 Š−CYN U −½15 Nx Š−CYN ⋅100

0

C15H22O714N5S

416.1236

0.29

100.0

1 2 3 4 5

C15H22O714N415N1S C15H22O714N315N2S C15H22O714N215N3S C15H22O714N115N4S C15H22O715N5S

417.1205 418.1178 419.1146 420.1116 421.1086

0.08 0.60 0.10 0.04 0.05

58.6 34.1 9.5 2.1 0.7

n=3) was spiked with CYN ISTD yielding a final concentration in the extract of approximately 15.0 μg/L U-[15N5]-CYN. Spiking experiments were conducted using samples from control plants. Lyophilized leaves of kale and vegetable mustard, 40 mg each (n=3), were spiked gravimetrically with CYN STD and ISTD yielding dry weight (DW)-based native CYN mass fractions of 200.0, 100.0, and 50.0 μg/kg DW representing realistic mass fraction as expected from a prior study [7]. The final U-[15N5]-CYN ISTD concentration in the plant extract was approximately 15.0 μg/L. Naturally contaminated and spiked vegetable plant samples were equilibrated with the standard spikes for at least 0.5 h. Afterwards, the samples were extracted by addition of 0.5 mL water and shaking the samples for 2 h at 50 °C using a thermomixer set to a horizontal shaking frequency of 700 min-1 (HLC BioTech—Ditabis AG, Pforzheim, Germany). The samples were centrifuged (14,000×g); the supernatant was loaded onto rinsed (6 mL MeOH/TFA (99/1; v/v) freshly prepared) and preconditioned (9 mL ultrapure water) selfpacked SPE cartridges (sorbent 90 mg GCB). The cartridges were washed by adding 6 mL of water and brought to dryness by applying a vacuum. CYN was eluted using 6 mL MeOH/TFA (99/1; v/v). After evaporating the eluent to dryness using a vacuum concentrator, the residue was dissolved in 500 μL water for analysis by LC-MS.

LC-MS/MS Analyses were performed using a 1100 Series HPLC system from Agilent Technologies (Waldbronn, Germany). For chromatography, a 2.1×150-mm, particle size 3 μm Atlantis C18 column with guard protection (Waters, Eschborn, Germany) was used. Eluent A was 98 % water, 1 % ACN, and 1 % MeOH and eluent B 40 % water, 30 % ACN, and 30 % MeOH, both eluents containing 5 mM ammonium acetate and 0.1 % (v/v) acetic acid. For freshwater samples and U-[15N5]-CYN confirmation, a gradient elution was used, starting with 100 % eluent A, decreasing eluent A to 0 % within 7 min, and then holding 0 % A for 3 min followed by equilibration of the column with eluent A for 8 min. Because of interfering matrix peaks, the gradient was adjusted for

%

vegetable plant samples, starting with 100 % eluent A for 3.5 min, decreasing eluent A to 0 % within 0.5 min, and then holding 0 % A for 4 min which was followed by increasing eluent A within 1 min and the equilibration of the column for 6.5 min. For all samples, a flow rate of 0.3 mL/min was applied and 10 μL were injected. Electrospray ionization (ESI) MS/MS data were acquired on an API 4000 triple-quadrupole MS/MS system (AB Sciex, Foster City, USA). The mass spectrometer operated with the source parameters as previously reported [7]. SRM transitions monitored for native CYN were 416.1→194.1 (quantifier, declustering potential (DP) 55 V, collision energy (CE) 49 V) and 416.1→336.1 (qualifier, DP 55 V, CE 31 V) and for U-[15N5]-CYN 421.1→197.1 (quantifier, DP 55 V, CE 49 V). Product ion spectra of m/z 416 and m/z 421 were acquired for a mass range of m/z 150–450 using a CE of 30 V. Data acquisition was done using Analyst 1.6.2 software (AB Sciex, Foster City, USA).

Liquid chromatography high-resolution mass spectrometry (LC-HRMS) ESI-HRMS data were acquired on an Ultimate 3000 RSLC nano-LC-HRMS system (Dionex, Amsterdam, The Netherlands) connected to an Orbitrap Exactive system (Thermo Fisher, Bremen, Germany). U-[15N5]-CYN dissolved in water (c=0.1 μg/mL) was injected to a 0.3×150-mm, particle size 2 μm, pore size 100 Å Acclaim PepMap RSLC C18 column with guard protection (Dionex, Amsterdam, The Netherlands). Mobile phase A was 100 % water and mobile phase B 80 % ACN and 20 % water, both acidified with 0.1 % formic acid (v/v). The gradient started with 98 % A, which was held for 2 min, and then increased to 95 % within 11 min at a flow rate of 4 μL/min. The injection volume was 10 μL. The mass spectrometer was equipped with an ESI source and operated at +5,000 V in positive ionization mode. Tube lens voltage was set to 130 V, and capillary voltage was set to 30 V. A capillary temperature of 275 °C was used. A mass range between 400 and 500 Da was recorded using a mass resolution of 100,000. Data were acquired with the Xcalibur 2.2 software (Thermo Fisher, Bremen, Germany).

15

N-labeled cylindrospermopsin and its application

5769

Biosynthesis of isotopically labeled CYN

formation of higher 15N-labeled CYN molecules is shown in Fig. 2 for three time points. Possible explanations for the slow replacement of 14 N atoms by 15N atoms in CYN are the long-lasting turnover to 14N-free medium and the incorporation of still remaining 14 N-containing compounds from the cyanobacterial biomass, e.g., glycine and arginine, into CYN [27]. The cultivated cyanobacterial strain is known to be able to fix nitrogen. However, nitrogen fixation is repressed in the presence of a rich source of combined nitrogen that was also offered by the medium in our experiments [28]. Nevertheless, biosynthesis of uniformly 15N-labeled CYN is limited due to isotopic purity of Na15NO3 applied for APH culture which was at least 98 at.% 15N. The formation of U-[15N 5]-CYN was successfully confirmed applying high-resolution and tandem mass spectrometric techniques. The analysis of U-[ 15 N 5]CYN by LC-MS/MS showed that U-[15N5]-CYN and native CYN have the same RTs. By recording product ion spectra, complementary fragmentation patterns were detected for both compounds (Fig. 3). According to the number of 15 N atoms present in the molecule, the m/z values of the respective fragment ions are increased. This correlation can also be observed for all stages of 15 N incorporation in the CYN molecule by LC-HRMS. The presence of all labeling stages was confirmed with mass errors less than 1 ppm (Table 1). A comparison between the different N-isotopic compositions of CYN showed that U-[15N5]-CYN is most appropriate to be applied as ISTD in SIDA. Native CYN represents less than 1 % in relation to U-[15N5]CYN (Table 1 and Supplementary Fig. S1). Therefore, the contribution of native CYN caused by commonly used quantities of ISTD for mass spectrometric purposes will be negligible. Furthermore, U-[15N 5]-CYN possesses the highest mass difference compared to native

Biosynthesis of U-[15N5]-CYN (450 μg) was achieved under defined medium conditions for CYN producing cyanobacteria. The dominating formation of uniformly 15N-labeled CYN was confirmed by LC-MS/MS, i.e., comparing the retention time (RT) and the fragmentation pattern to a native CYN STD and by determining the exact mass using LCHRMS analysis. CYN isolated from APH grown on common Z medium contains 77 % U-[14N5]-CYN and 0.003 % U-[15N5]-CYN referred to the sum of all stages of 15N incorporation. After 18 months, growing cyanobacteria in 15N-containing medium nearly native CYN-free standard was obtained. The native CYN proportion accounts for less than 1 % of the sum of all stages of 15N incorporation, whereas the U-[15N5]-CYN proportion accounts for more than 65 %. The time course of the

Fig. 2 N-isotopic pattern for CYN when growing cyanobacteria in Na14NO3-containing medium (0 months) and after cultivation in Na15NO3-containing medium for 2 months and for 18 months

Calibration and quantification Stock solution of CYN was prepared in ultrapure water and stored at −20 °C. The concentration of the CYN stock solution was determined by using the certified CYN standard. Fivepoint equidistant calibrations were prepared in the ranges of 2.5–20.0 μg/L and 20.0–100.0 μg/L for CYN quantification in freshwater and vegetable plants by SIDA. Therefore, the respective amounts of CYN and a constant amount of U-[15N5]-CYN ISTD (resulting in a final concentration of 30 μg/L) were weight into HPLC vials. For evaluation of the potential of the SIDA method to compensate for matrix effects, a seven-point matrix calibration in the range of 5.0–100.0 μg/L was prepared. Variable amounts of CYN and constant amounts of ISTD (10.0 μg/L final content) were weighted into 1.5-mL reaction tubes followed by solvent evaporation using a vacuum concentrator. The residue was dissolved in uncontaminated kale, vegetable mustard, or freshwater (concentration factor 1,000) matrix extract. The extracts were obtained following the same procedure as described for vegetable plant and freshwater sample preparation. Quantification of CYN was realized by calculating the peak area ratio of the STD compared to the ISTD and using calibration curves obtained by simple linear regression as described by Asam et al. [26]. The limit of detection (LOD) determined for CYN is based on a signal-to-noise ratio (S/N) of 3 for CYN. The LODs for CYN in freshwater concentrated by a factor of 1,000 and in vegetable plant leaves are 4.5 ng/L and 15 μg/kg dry weight, respectively. The LOQ is expressed as three times the LOD.

Results and discussion

5770

K. Kittler et al.

Fig. 3 Structure and main fragments of (a) U-[14N5]-CYN and (b) U-[15N5]-CYN in ESIpositive mode and the corresponding MS2 product ion spectra

CYN leading to the least interferences with heavy isotope containing native CYN species.

calculation of the CYN amount. The RSDs are caused by the changing ionization of the analyte within one set of LC-MS/MS measurements. Moreover, uncontaminated

U-[15N5]-CYN in LC-MS/MS analysis In order to evaluate the suitability of the biosynthesized labeled standard for the compensation of matrix effects in LC-MS/MS analysis, matrix calibration curves with and without U-[15N5]-CYN ISTD correction were compared. To focus on the LC-MS/MS analysis step, CYN was spiked after the SPE procedure and before measurement to CYN-free matrices. Figure 4 shows the obtained calibration curves. The CYN signal intensity is dependent on the type of matrix which is demonstrated by varying slopes of the calibration curves up to a factor of 10 (Fig. 4a). Varying slopes are observable for different matrices intra-day as well as for same matrices inter-day. Ultrapure water I and II represent measurements from identical matrix however recorded on different days. In contrast, congruent regression curves were obtained for the tested matrices when using the SIDA technique (Fig. 4b). To some extent, matrix effects can also be compensated by matrix calibration. However, if the matrix calibration approach was applied, the signal intensity of CYN deviates substantially for each individual calibration level (see Fig. 4a). The high mean relative standard deviation (RSD) of 26 % leads to a nonprecise

Fig. 4 Calibrations for CYN (STD) in ultrapure water, freshwater, vegetable mustard, and kale (a) without and (b) with U-[15N5]-CYN internal standard (ISTD) correction. STD and ISTD were spiked after cleanup prior to LC-MS/MS analysis. Data are reported as mean±SD (n=3)

15

N-labeled cylindrospermopsin and its application

5771

matrix is not always available, and a matrix-matched calibration needs to be prepared for each matrix separately. Thus, application of SIDA represents a much more versatile approach. The calibrations utilizing the biosynthesized U-[15N5]CYN ISTD are independent of the matrix applied which is demonstrated by the congruent regression curves. Additionally, the mean RSD of the SIDA calibrations was reduced to approximately 6 % for all individual calibration levels compared to the mean RSD calculated for matrix calibrations.

CYN quantification by SIDA To evaluate the entire developed analytical process, which includes sample extraction, cleanup, and LCMS/MS analysis, recoveries and RSDs for spiking experiments were determined by applying SIDA. For this purpose, CYN STD and ISTD were spiked to CYN-free freshwater, kale, and vegetable mustard leaves prior to sample treatment. The mean recovery rates and the RSDs of CYN obtained from the different spiking levels are summarized in Table 2.

Vegetable plant leaves The recovered CYN amount in kale and vegetable mustard leaves ranged between 98 and 103 % with RSDs not exceeding 5 %. CYN may bind easily to tissues because of its structural features [29, 30]. Although vegetable plant matrix is highly complex containing proteins, carbohydrates, and fats, CYN was apparently completely recovered. Thus, binding effects were satisfactorily compensated. Despite the accurate results, quantification of CYN in vegetable plants was more challenging than in freshwater (Fig. 5). Components with CYN-like SRM transitions are present in both Brassica matrices, requiring an optimized chromatographic HPLC system to separate CYN from the interfering matrix. The sample cleanup had no impact on S/N. Matrix with CYN-like SRM transitions was not significantly removed. Although the applied SPE method resulted in CYN recoveries of approximately 95 % without ISTD correction, CYN signal intensity was not increased. However, different from raw extract analysis, the analysis of purified samples did not lead to a significant decrease of overall sensitivity of MS measurements within several sequences. In summary, the developed method is suitable for CYN quantification in Brassica leaves. Naturally contaminated samples

Freshwater Recoveries from 99 to 101 % and small RSDs were obtained over a wide concentration range when CYN was quantified in concentrated freshwater. In general, CYN losses could be expected if SPE is applied for freshwater concentration especially in combination with samples with a high dissolved organic carbon content [21]. Nevertheless, SPE is mandatory to detect low CYN concentrations. The developed method allows for a reliable CYN determination in natural waters up to a factor of 50 below the proposed guideline safety value of 1 μg/L.

Table 2 Determined CYN recoveries and RSDs in various matrices. The spiked CYN concentration is given for freshwater. The spiked CYN mass fraction refers to vegetable leaf dry weight (DW). CYN was quantified by SIDA based on calibration in ultrapure water. Data are reported as mean±RSD (n=3)

The developed SIDA for quantification of CYN was applied to the analysis of naturally CYN contaminated matrices. Seven different freshwater bodies, irrigation water for plant exposure experiments, and leaf materials of two different Brassica species were analyzed (Table 3). Freshwater In two out of seven freshwater body samples, CYN was detectable as well as quantifiable; none of these two samples exceed the guideline safety value of 1 μg/L CYN for drinking water. The present study exclusively considered extracellular

Matrix

Spiked CYN amount

Recovery±RSD [%]

Concentration factor

Freshwater

0.03 μg/L 1.0 μg/L 10.0 μg/L 50.0 μg/kg DW 100.0 μg/kg DW 200.0 μg/kg DW 50.0 μg/kg DW 100.0 μg/kg DW 200.0 μg/kg DW

101.4±1.3 100.0±2.1 99.0±1.3 100.2±3.8 100.4±3.5 102.5±3.1 101.0±4.0 97.8±2.3 99.3±2.1

1,000 100 10 – – – – – –

Kale

Vegetable mustard

5772

K. Kittler et al. Table 3 SIDA-LC-MS/MS measurement of CYN in freshwater and vegetable leaf matrices. CYN concentration is given for freshwater and CYN mass fraction is calculated for vegetable leaf dry weight (DW). Mean and RSD are given for n=3 samples

Fig. 5 Overlay of LC-ESI-MS/MS chromatograms for the SRM transition 416→194 with and without sample cleanup. Kale was spiked with CYN (50 μg/kg) to investigate the influence of sample cleanup

CYN and not bound CYN. However, Rücker et al. stated that the dissolved CYN fraction accounts for up to 100 % of total CYN fraction [4]. Thus, measuring the dissolved CYN is sufficient for an initial screening.

Vegetable plant leaves CYN was quantifiable in the leaves of kale and vegetable mustard when only the roots were irrigated with water containing CYN providing proof for the systemic availability of CYN to Brassica species. In average, the CYN content in vegetable mustard (23.9±11.9 μg/kg fresh weight) was higher than in kale (15.0±1.4 μg/kg fresh weight). Due to high intersample variation for vegetable mustard, this finding is statistically not significant. Nevertheless, the CYN content was precisely quantified for individual samples as it can be seen from Table 3. The water applied for irrigation of the test plants contained 74.7 μg/L CYN, representing a concentration which is four times higher than in a previous experiment [7]. Corresponding to the increased exposure level, also an approximately fourfold increase of CYN content in the leaf fresh weight is observable. This linear increase of CYN content in the leaves is in agreement with the findings for Sinapis alba seedlings [7]. However, the almost linear increase of systemic CYN availability is accompanied by a high inter-individual variation. Taken all together, the results demonstrate the successful biosynthesis of U-[15N5]-CYN and development of a SIDA-LC-MS/MS method for a reliable and accurate quantification of CYN in freshwater and vegetable plants. Applying the developed SIDA-LC-MS/MS method, CYN recoveries of approximately 100 % and RSDs not exceeding 5 % were achieved, which conforms to the provisions of EU Regulation 657/2002 [31]. Due to the lack of authentic reference material with known CYN content, the recovery was investigated for spiked samples.

Sample

CYN amount

RSD [%]

Irrigation water for plant exposure experiments “Rangsdorfer See” (Germany) “Motzener See” (Germany) “Seddiner See” (Germany) “Peene-Strom” (Germany) “Pond Juja” (Kenya) “Pond Juja” irrigation water (Kenya)

74.7 μg/L

0.8

n.d. 60.8 ng/L 17.9 ng/L n.d. n.d. n.d.

Local water supply (Kenya) Vegetable mustard 1 Vegetable mustard 2 Vegetable mustard 3 Kale 1 Kale 2 Kale 3

n.d. 480.6 327.1 150.9 184.0 180.8 155.0

μg/kg DW μg/kg DW μg/kg DW μg/kg DW μg/kg DW μg/kg DW

1.3 2.5

1.5 2.0 1.9 0.4 1.8 3.0

Therefore, an underestimation of CYN is possible for real samples containing CYN in chemically bound form which cannot be entirely compensated by addition of the U-[15N5]-CYN ISTD. However, the newly developed isotopically labeled standard could compensate for the matrix-dependent response of CYN in LC-MS/MS analyses as well as for analyte losses due to matrix binding and SPE cleanup. Even for trace amounts of CYN and in complex matrices, the reliable quantification is feasible by SIDA which is essential for a comprehensive environmental monitoring. Recent reports on the widespread occurrence of CYN as well as the identification of irrigated plants as an additional exposure route underline the relevance of reliable CYN analysis. Development of SIDA for food of aquatic origin, representing another important human exposure source, was not part of the present study and should be part of future research. Furthermore, as a reference method for the development of immunological tests that would enable a quick and reliable on-site CYN determination or in toxicological investigations requiring accurate and sensitive CYN quantification, SIDA using U-[15N5]-CYN will be a helpful tool.

Acknowledgments We would like to thank Elke Büsch for plant culture as well as Wolfgang Pritzkow and Shireen Weise for technical assistance with the Orbitrap Exactive system. We greatly thank Jutta Fastner and Karina Preussel from the UBA for providing the cyanobacterial strain A. flos-aquae 22D11.

15

N-labeled cylindrospermopsin and its application

References 1. Hawkins PR, Runnegar MT, Jackson AR, Falconer IR (1985) Severe hepatotoxicity caused by the tropical cyanobacterium (blue-green alga) Cylindrospermopsis raciborskii (Woloszynska) Seenaya and Subba Raju isolated from a domestic water supply reservoir. Appl Environ Microbiol 50(5):1292–1295 2. Kinnear S (2010) Cylindrospermopsin: a decade of progress on bioaccumulation research. Marine Drugs 8(3):542–564. doi:10. 3390/Md8030542 3. Saker ML, Eaglesham GK (1999) The accumulation of cylindrospermopsin from the cyanobacterium Cylindrospermopsis raciborskii in tissues of the Redclaw crayfish Cherax quadricarinatus. Toxicon 37(7):1065–1077 4. Rucker J, Stuken A, Nixdorf B, Fastner J, Chorus I, Wiedner C (2007) Concentrations of particulate and dissolved cylindrospermopsin in 21 Aphanizomenon-dominated temperate lakes. Toxicon 50(6):800–809. doi:10.1016/j.toxicon. 2007.06.019 5. Saker ML, Metcalf JS, Codd GA, Vasconcelos VM (2004) Accumulation and depuration of the cyanobacterial toxin cylindrospermopsin in the freshwater mussel Anodonta cygnea. Toxicon 43(2):185–194. doi:10.1016/j.toxicon.2003.11.022 6. Gutierrez-Praena D, Jos A, Pichardo S, Moreno IM, Camean AM (2013) Presence and bioaccumulation of microcystins and cylindrospermopsin in food and the effectiveness of some cooking techniques at decreasing their concentrations: A review. Food Chem Toxicol 53:139–152. doi:10.1016/j.fct.2012.10.062 7. Kittler K, Schreiner M, Krumbein A, Manzei S, Koch M, Rohn S, Maul R (2012) Uptake of the cyanobacterial toxin cylindrospermopsin in Brassica vegetables. Food Chem 133(3):875–879. doi:10.1016/j. foodchem.2012.01.107 8. Humpage AR, Falconer IR (2003) Oral toxicity of the cyanobacterial toxin cylindrospermopsin in male Swiss albino mice: determination of no observed adverse effect level for deriving a drinking water guideline value. Environ Toxicol 18(2):94– 103. doi:10.1002/tox.10104 9. Ministry of Health (2008) Drinking-water standards for New Zealand 2005 (revised 2008). Water Quality Standards, Ministry of Health, Wellington, New Zealand 10. Saker ML, Nogueira ICG, Vasconcelos VM, Neilan BA, Eaglesham GK, Pereira P (2003) First report and toxicological assessment of the cyanobacterium Cylindrospermopsis raciborskii from Portuguese freshwaters. Ecotoxicology and Environmental Safety 55(2):243– 250. doi:10.1016/S0147-6513(02)00043-X 11. Elliott CT, Redshaw CH, George SE, Campbell K (2013) First development and characterisation of polyclonal and monoclonal antibodies to the emerging fresh water toxin cylindrospermopsin. Harmful Algae 24:10–19. doi:10.1016/j.hal.2012.12.005 12. Harada K, Ohtani I, Iwamoto K, Suzuki M, Watanabe MF, Watanabe M, Terao K (1994) Isolation of cylindrospermopsin from a cyanobacterium Umezakia natans and its screening method. Toxicon 32(1): 73–84. doi:10.1016/0041-0101(94)90023-X 13. Froscio SM, Humpage AR, Burcham PC, Falconer IR (2001) Cellfree protein synthesis inhibition assay for the cyanobacterial toxin cylindrospermopsin. Environ Toxicol 16(5):408–412 14. Guzman-Guillen R, Prieto AI, Gonzalez AG, Soria-Diaz ME, Camean AM (2012) Cylindrospermopsin determination in water by LC-MS/MS: optimization and validation of the method and application to real samples. Environ Toxicol Chem 31(10):2233–2238. doi: 10.1002/etc.1954

5773 15. Gallo P, Fabbrocino S, Cerulo MG, Ferranti P, Bruno M, Serpe L (2009) Determination of cylindrospermopsin in freshwaters and fish tissue by liquid chromatography coupled to electrospray ion trap mass spectrometry. Rapid Commun Mass Spectrom 23(20):3279– 3284. doi:10.1002/rcm.4243 16. van Apeldoorn ME, van Egmond HP, Speijers GJ, Bakker GJ (2007) Toxins of cyanobacteria. Mol Nutr Food Res 51(1):7–60. doi:10. 1002/mnfr.200600185 17. Asam S, Konitzer K, Schieberle P, Rychlik M (2009) Stable isotope dilution assays of alternariol and alternariol monomethyl ether in beverages. J Agric Food Chem 57(12):5152–5160. doi:10.1021/ jf900450w 18. Sano T, Takagi H, Nagano K, Nishikawa M, Kaya K (2011) Accurate LC-MS analyses for microcystins using per-N-15-labeled microcystins. Anal Bioanal Chem 399(7):2511–2516. doi:10.1007/ s00216-010-4639-y 19. Yen HK, Lin TF, Liao PC (2011) Simultaneous detection of nine cyanotoxins in drinking water using dual solid-phase extraction and liquid chromatography-mass spectrometry. Toxicon 58(2):209–218. doi:10.1016/j.toxicon.2011.06.003 20. Metcalf JS, Beattie KA, Saker ML, Codd GA (2002) Effects of organic solvents on the high performance liquid chromatographic analysis of the cyanobacterial toxin cylindrospermopsin and its recovery from environmental eutrophic waters by solid phase extraction. FEMS Microbiol Lett 216(2):159–164 21. Wormer L, Carrasco D, Cires S, Quesada A (2009) Advances in solid phase extraction of the cyanobacterial toxin cylindrospermopsin. Limnol Oceanogr-Meth 7:568–575 22. Kikuchi S, Kubo T, Kaya K (2007) Cylindrospermopsin determination using 2-[4-(2-hydroxyethyl)-1-piperazinyl]ethanesulfonic acid (HEPES) as the internal standard. Anal Chim Acta 583(1):124–127. doi:10.1016/j.aca.2006.10.007 23. Preussel K, Stuken A, Wiedner C, Chorus I, Fastner J (2006) First report on cylindrospermopsin producing Aphanizomenon flos-aquae (Cyanobacteria) isolated from two German lakes. Toxicon 47(2): 156–162. doi:10.1016/j.toxicon.2005.10.013 24. Falch BS, Konig GM, Wright AD, Sticher O, Angerhofer CK, Pezzuto JM, Bachmann H (1995) Biological activities of cyanobacteria: evaluation of extracts and pure compounds. Planta Med 61(4):321–328. doi:10.1055/s-2006-958092 25. Kuhl A, Lorenzen H (1964) Handling and culturing of Chlorella. In: Preston DM (ed) Methods of cell physiology, vol 1. Academic, London, pp 159–187 26. Asam S, Liu Y, Konitzer K, Rychlik M (2011) Development of a stable isotope dilution assay for tenuazonic acid. J Agric Food Chem 59(7):2980–2987. doi:10.1021/jf104270e 27. Mihali TK, Kellmann R, Muenchhoff J, Barrow KD, Neilan BA (2008) Characterization of the gene cluster responsible for cylindrospermopsin biosynthesis. Appl Environ Microbiol 74(3): 716–722. doi:10.1128/AEM.01988-07 28. Adams DG (2000) Heterocyst formation in cyanobacteria. Curr Opin Microbiol 3(6):618–624 29. Duy TN, Lam PK, Shaw GR, Connell DW (2000) Toxicology and risk assessment of freshwater cyanobacterial (blue-green algal) toxins in water. Rev Environ Contam Toxicol 163:113–185 30. Froscio SM, Humpage AR, Wickramasinghe W, Shaw G, Falconer IR (2008) Interaction of the cyanobacterial toxin cylindrospermopsin with the eukaryotic protein synthesis system. Toxicon 51(2):191– 198. doi:10.1016/j.toxicon.2007.09.001 31. Commission Regulation (2002) No 657/2002 Implementing Council Directive 96/23/EC concerning the performance of analytical methods and the interpretation of results. Off J Eur Union L71

5774

K. Kittler et al. Katrin Kittler has been working on the investigation of freshwater and marine biotoxins in different matrices and studied their metabolism and toxicity during her PhD research at the Federal Institute for Materials Research and Testing (BAM), Germany. She developed and applied different cell biological, microbiological, and analytical techniques.

Matthias Koch is Head of the Division of Food Analysis at the Federal Institute for Materials Research and Testing (BAM). He has been working for several years on the development of analytical methods and certified reference materials for organic contaminants (in particular mycotoxins) in food.

Holger Hoffmann Is a student of the Humboldt University Berlin and completed his diploma thesis at the Federal Institute for Materials Research and Testing (BAM) in the Division of Food Analysis. He is presently working as a PhD student in the Division of Immunoanalytics. His topics are ELISA, LC-UV, LC-MS/MS, and HRMS.

Sascha Rohn is a Full Professor for Food Chemistry at the University of Hamburg, Hamburg School of Food Science, Germany. His group deals with the analysis of secondary plant metabolites and their antioxidant activity, but especially with the reactivity and stability of these bioactive compounds. The aim is to identify degradation products that serve as process markers during food/feed processing or as biomarkers in nutritional physiology.

Franziska Lindemann Is employed at the Federal Institute for Materials Research and Testing (BAM), Germany, and was responsible for the isolation and analysis of bioactive compounds in different matrices by application of liquid chromatographic analytical techniques.

Ronald Maul is a senior scientist at the Leibniz-Institute of Vegetable and Ornamental Crops in Großbeeren (IGZ). In the Quality Department, he investigates secondary metabolites and natural contaminants in plants and plantbased food. Prior to this, he spent several years at the Federal Institute for Material Research and Testing (BAM) establishing HPLC-MS/MS-based methods for the analysis of microbial toxins in biological matrices.