A
GENERIC
METHOD
FOR
REGULATED
AND
UNREGULATED
PHYCOTOXINS
IN
VARIOUS
MATRICES
WITH
LC-HRMS
3
MSc Chemistry
Analytical sciences
Master Thesis
A generic method for regulated and unregulated
phycotoxins in various matrices with LC-hrMS
by
M. D. Klijnstra
April 2016
Supervisor:
dr W. Th. Kok
Daily Supervisor:
dr A. Gerssen
RIKILT Wageningen UR
5
A
BSTRACT
Phycotoxins such as marine biotoxins and cyanotoxins are produced by certain algae and cyanobacteria which are naturally occurring in marine and fresh waters. Phycotoxins can accumulate in various marine and fresh water species such as fish, crabs or shellfish (mussels, oysters, scallops and clams). By processing toxin producing algae, contaminated fish or shellfish, phycotoxins can also end up in food or food supplements. When these contaminated products are consumed or when there is contact with certain toxins, e.g., in swimming water, they may cause severe intoxication symptoms, such as skin irritation, paralysis, diarrhoea and even death may occur. Every year phycotoxins are held responsible for approximately 60,000 human intoxications.
Legislation has been developed and monitoring programs have been established worldwide in order to prevent intoxication of the consumer. Analytical methods are available to analyse regulated phycotoxins. However, these methods are only suitable for a small specific group of toxins and/or a specific matrix (mainly shellfish). Due to their different chemical properties it is difficult to analyse all phycotoxins in one single analytical method. For instance, lipophilic compounds dissolve more easily in organic solvents and can be separated on reversed phase liquid chromatography (LC) columns. Hydrophilic compounds dissolve best in water and can be retained on hydrophilic interaction LC columns (HILIC).
For cases in which symptoms cannot be directly related to specific regulated phycotoxins, a screening method is required. In this work such a screening method was developed and validated for all types of phycotoxins in different matrices such as tissue, fresh and sea water and food supplements. Two LC methods were developed to analyse sample extracts; one for hydrophilic and one for lipophilic phycotoxins. Sample extracts were measured in full scan mode with an Orbitrap high resolution mass spectrometry (hrMS). Additionally, a database was created to process the data. Unlike various other modes of MS/MS acquisition, LC-hrMS allows untargeted measurements with the possibility to detect additional compounds, which were not foreseen to be of interest at the time of the measurement, retrospectively.
The validation of the screening method for tissue samples was successful. Furthermore, it was shown that regulated lipophilic phycotoxins, domoic acid and some paralytic shellfish poisoning (PSP) toxins can be quantified in shellfish at 0.5 or 1 times the permitted level. However, some PSPs gave poor peak shapes which caused difficulties during processing of the results. The validation of the screening method for water samples was also successful, except for hydrophilic phycotoxins in sea water. During validation it appeared that the method for sea water had to be modified slightly due to problems with the salt content. Recoveries of lipophilic phycotoxins spiked to food supplements ranged from 9 to 102%, depending on matrix effects from sample to sample. Therefore it was difficult to validate the method for lipophilic phycotoxins in food supplements. When phycotoxins are found based on the library and confirmed with a fragment ion using the screening method, it is still a tentative confirmation because fragment ions are analysed over a range of precursor ions and retention times are often unknown. To confirm the presence of phycotoxins a standard is needed, or NMR analysis needs to be done. For most phycotoxins for which standards are available confirmation methods already exist. For the palytoxin-group toxins (PlTXs) an analytical method was not yet available and therefore an LC-MS/MS method was established for the quantitation and confirmation of PlTXs. However, high quantities of PlTXs were necessary to detect the compounds with the screening method, since the spectrum is complicated (multiple precursor ions), and therefore a poor sensitivity was obtained. Within the confirmation method cleavage fragments from an oxidative fragmentation reaction are included which solves any sensitivity issues for PlTXs. However, the oxidation step was too specific to be included in the screening method.
7
C
ONTENTS
1.
Introduction ... 9
2.
Materials and methods ... 13
2.1
Chemicals and standards ... 13
2.2
Preparation of standard solutions ... 15
2.3
Preparation of extracts ... 15
2.4
Instrumentation ... 17
2.5
Validation ... 19
3.
Results and discussion ... 21
3.1
Optimization of extraction procedures ... 21
3.2
Chromatography... 26
3.3
MS measurements... 30
3.3
Validation ... 30
3.4
Confirmation of palytoxin-group toxins by LC-MS/MS ... 33
4.
Conclusion ... 37
5.
Acknowledgement ... 38
6.
References ... 39
Appendices ... 43
Appendix 1: Structures ... 44
Appendix 2: Database (part as example) ... 53
Appendix 3: RIKILT Standards operating procedures phycotoxins ... 57
9
1. I
NTRODUCTION
Phycotoxins such as marine biotoxins and cyanotoxins are produced by certain algae and cyanobacteria that are naturally occurring in marine and fresh waters. When a population of algae is rapidly increasing or accumulating in a water system this is called an algal bloom. An algal bloom that has negative effects for other organisms (e.g., via production of phycotoxins or by shading) is called a harmful algal blooms (HAB).
HABs are caused by the presence of high nutrient concentrations in the water by influx through river deltas and rainfall and HAB formation can be enhanced by other weather factors such as wind and temperature [1]. When phycotoxins are present they can accumulate in various marine species such as fish, crabs or shellfish (mussels, oysters, scallops and clams). In shellfish, toxins accumulate mainly in the digestive glands without causing intoxication to the shellfish itself [2]. Various shellfish species can metabolize some phycotoxins into fatty acid esters, which can cause a delayed onset of toxic symptoms. This esterification is known for brevetoxins (PbTxs), cyclic imines (CIs), pectenotoxins (PTXs), okadaic acid (OA) and dinophysistoxins (DTXs) [3-8]. Furthermore yessotoxins (YTXs) and PTXs can be metabolized by shellfish to analogues [9].
By processing toxin producing algae, contaminated fish or shellfish, phycotoxins can also end up in food or food supplements. When these contaminated products are consumed or when there is contact with certain toxins, e.g., in swimming water, they may cause severe intoxication symptoms, such as skin irritation, paralysis, diarrhoea and even death may occur [10, 11]. In the past intoxications have been reported [12-14] and throughout the world, phycotoxins produced by algae are held responsible for approximately 60,000 human intoxications every year [15].
Some phycotoxins are classified by the syndromes they cause: paralytic shellfish poisoning (PSP) is caused by saxitoxin (STX) and analogues, diarrhetic shellfish poisoning (DSP) is caused by OA and DTXs, amnesic shellfish poisoning (ASP) is caused by domoic acid (DA), and neurotoxic shellfish poisoning (NSP) is caused by PbTxs [16]. However, there are many more phycotoxins known, including: azaspiracids (AZAs), YTXs, CIs, PTXs, microcystins and nodularins (MCs), cylindrospermopsins (CYNs), anatoxins (ATXs), tetrodotoxins (TTXs), ciguatoxins (CTXs) and palytoxins (PlTXs). AZAs cause diarrhoea; PTXs, CYNs and MCs are hepatotoxic; CIs, ATXs, and TTXs are neurotoxic; CTXs show gastrointestinal and neurological symptoms and PlTXs cause gastrointestinal problems or respiratory distress [17-22]. When human are exposed to high concentrations, some phycotoxins i.e. TTXs and PbTxs can be lethal. YTXs are lethal to mice after intraperitoneal injection but not for human after consumption of contaminated shellfish [23]. All mentioned toxin groups contain multiple analogues; the largest group are the MCs with more than 160 different structures reported in literature and registered in Scifinder [24, 25]. Some analogues are more toxic than others and therefore for some phycotoxin groups toxicity equivalent factors (TEF) have been assigned.
Less than ten percent of all phycotoxins described in literature are available as a (certified) standard. Standards are isolated from contaminated shellfish or algae and there are not more than a few producers worldwide that produce certified phycotoxin standards. Furthermore the availability of contaminated shellfish and algae is limited and it is time-consuming to produce such purified standards, and therefore standards are relatively expensive [26].
Appendix 1 gives an overview of abbreviations and structures of all available standards. Molecular masses vary from 118 Da to 3380 Da and phycotoxins can be divided into two classes based on their water solubility. In general all phycotoxins with a molecular mass below 500 Da are considered to be hydrophilic phycotoxins and all phycotoxins with a mass above 500 Da are considered to be lipophilic. The lipophilic phycotoxins include compounds such as OA, DTXs, AZAs, PTXs, and YTXs. The hydrophilic phycotoxins include DA, STXs and CYNs.
10 The analysis of lipophilic and hydrophilic compounds is very different. Lipophilic compounds tend to dissolve better in organic solvents and are separated on reversed phase LC-columns. Hydrophilic compounds dissolve better in water and are preferably separated with Hydrophilic Interaction Liquid Chromatography (HILIC) columns.
Legislation and monitoring programs are established worldwide in order to prevent intoxication of shellfish consumers [27, 28]. However there are some differences within legislation methods applied around the world. Table 1 gives an overview of legislation in the European Union, CODEX guidelines and opinions of the European Food Safety Authority (EFSA) on their maximum allowed concentrations. These levels are only applicable to fish or shellfish and they are the sum of the concentrations of all analogues taking TEFs into account. Legislation in the European Union is comparable with CODEX guidelines except for PTXs and YTXs, which are additional, and PbTxs, as PbTxs do not occur in European waters. However, EFSA opinions are very different from the EU and CODEX, where for most groups toxic levels are estimated lower by EFSA. This could be a concern as most of the recommended levels by EFSA are allowed to be exceeded. For MCs there is a guideline for drinking water issued by the World Health Organisation of 1 µg L-1.
TABLE 1: LEGISLATION, GUIDELINES AND OPINIONS [29-40]
Toxin group EU (µg kg-1) CODEX (µg kg-1) EFSA opinions (µg kg-1) OA and DTXs (DSP) YES – 160 YES – 160 45
PTXs YES – 160* NO 120 YTXs YES – 3750 NO 3750 AZAs YES – 160 YES – 160 30 DA (ASP) YES – 20000 YES – 20000 4500 STXs (PSP) YES – 800 YES – 800 75
CIs NO NO More research needed
PlTXs NO NO 30
CTXs YES YES – 0.01 More research needed PbTxs (NSP) NO YES – 800 -
MCs NO NO -
TTXs NO NO -
ATXs NO NO -
CYNs NO NO -
* Included in the OA and DTX group
Besides variations in legislation there are also differences in the methods for analysis applied. Bioassays such as mouse and rat bioassays were common methods for the determination of phycotoxins in the past. With the official mouse bioassays for PSP and DSP toxins a mouse is injected intraperitoneally with a shellfish extract, when the mouse dies within 10 minutes (PSP) or 24 hours (DSP) the test result is positive [30, 33]. For the official rat bioassay for DPS toxins rats are starved for 24 hours and then fed with shellfish meat, the test result is positive when the rat is suffering from diarrhoea after 16 hours [30]. The disadvantages are that the response to possible phycotoxins might be specific for the animal and humans react different. Furthermore, the administration route of the mouse bioassay is different from normal consumption which makes it difficult to extrapolate results to human potency. Moreover, there is a growing resistance against the use of animals for experiments [41]. The main advantage of animal testing is that the selectivity is low; therefore the assay may also be sensitive towards possible unknown toxins. Until now, LC-MS/MS, HPLC-FLD and HPLC-UV have been shown to be the techniques with the highest sensitivities and selectivities [42] and for all regulated phycotoxins such analytical methods are available; however these methods are only suitable for a small specific group of toxins and/or a specific matrix (mainly shellfish). Moreover, these analytical methods do not detect any
11 phycotoxins that are not present during method development and included in the method. In general methods are only available for phycotoxins for which regulatory limits are established. Furthermore, limits of detection are for some methods above regulatory limits.
For PSP toxins the official inter-lab validated analytical methods are with pre- or post-column derivatisation of the sample extracts and analysis with fluorescence detection [33]. Liquid chromatography coupled to (triple quad) mass spectrometry (LC-MS/MS) methods for PSP toxins are described with HILIC separation, with various types of columns as zwitterion or amide based columns [43-45] and ion-pair LC-MS [46]. Furthermore TTXs can be separated with a HILIC column [47]. CYNs, further hydrophilic compounds, can be separated with a Zorbax Sb–Aq or a Luna PFP(2) column [48, 49]. In general HILIC methods are more sensitive to matrix effects which may cause shifts in retention time.
DSP toxins can be combined together with PTXs, YTXs, AZAs and CIs in one method for lipophilic toxins. Acidic, neutral or alkaline reversed phase chromatographic conditions are used in combination with a C8 or C18 column for separation of these lipophilic phycotoxins [42, 50]. When measured with MS, OA, DTXs and YTX are preferably analysed in the negative ionisation mode to obtain better sensitivity, and AZAs and CIs in the positive ionisation mode. The advantage of neutral and acidic conditions is that AZAs have better peak shapes as AZAs are zwitterions at a high pH; the advantage of alkaline conditions is that no polarity switch is needed to measure all compounds in the same run, because compounds analysed in the negative and positive mode are separated. Under acidic or neutral conditions the charge of the CIs is different and CIs elute more rapidly. When a conventional MS is used the disadvantage of polarity switching is that less data points can be generated which is compromising peak shapes and sensitivity. DSP toxins tend to form esters when present in shellfish. Many different esters, from C16 to C22 chains and with different saturations, are formed. They can be analysed intact [7], however, it is impossible to have standards for all of them and therefore esters can be transformed to the deconjugated form by alkaline hydrolysis [51].
Existing methods described for MCs are mostly suitable for water samples. MCs can be separated under acidic conditions in combination with a C18 column and are analysed in the positive ionisation mode [52]. PlTXs, PbTxs and CTXs are like MCs separated under acidic conditions in combination with a C18 column and are also analysed in the positive ionisation mode [53-55]. DA is retained on reversed phase columns as well as on HILIC columns; therefore DA is often included in multitoxin methods for DSP or PSP toxins and can be analysed in the positive and the negative ionisation mode [56, 57].
PlTXs is the group with the largest molecules. Because of the many functional groups at the molecule multiple adducts, charge states and losses of water are seen in the spectrum, which causes poor sensitivity. To obtain a better sensitivity for PlTXs an oxidative fragmentation reaction is described for reduction of molecules to specific cleavage fragments [58].
For all EU regulated phycotoxins or phycotoxins where an EFSA opinion is available, confirmative methods are available within RIKILT, except for PlTXs. For PlTXs an LC-MS/MS method needed to be developed, this was part of the research conducted within this thesis.
For untargeted analysis liquid chromatography combined with high resolution mass spectrometry (LC-hrMS) is used. This technique allows untargeted measurements with the possibility to detect additional compounds, which were not foreseen to be of interest at the time of the measurement, retrospectively, unlike triple quadrupole instruments operating in multiple reaction monitoring (MRM) mode. It has been shown that LC-hrMS is a useful technique to screen food samples for the presence of a high variety of analytes [59, 60]. For phycotoxins, only a few methods are described using hrMS, all suitable for only lipophilic toxins [61-63]. LC-hrMS makes it possible to approach food toxicant analysis in a different way with other possibilities; however, this brings also new challenges. The number of analytes that can be detected is too large to process and verify
12 manually as it is typically done for quantitative MS/MS methods. For this, databases with target analytes and sufficient information as retention or fragmentation information are needed to facilitate automated detection. When such a database is available, extraction of the analytes of interest from the raw data can be done automatically by software. There are several software packages to process data like MetAlign [64] which is developed at RIKILT or Tracefinder from Thermo Scientific. However, at this stage MetAlign cannot handle multiple charged fragments and Tracefinder is still under development because there are still some issues with the software to be solved.
In this study we have investigated the possibilities of LC-hrMS as a generic technique to screen for all kinds of phycotoxins in different matrices such as tissue, fresh and sea water and food supplements. This method will be applied in case of an incident where symptoms cannot be directly related to the regulated phycotoxins. The aim was to develop a method to extract different sample types with a generic extraction method and to analyse sample extracts with an LC-method for hydrophilic and one for lipophilic phycotoxins. In order to process the data a database needed to be created from literature and a protocol for data processing needed to be established. Furthermore, as mentioned before, for the PlTXs an LC-MS/MS confirmatory method needed to be developed.
13
2. M
ATERIALS AND METHODS
2.1
C
HEMICALS AND STANDARDS
Table 2 lists all chemicals used, abbreviations, concentration or purity and suppliers. Table 3 and Table 4 list all standards used divided into hydrophilic and lipophilic phycotoxins.
TABLE 2: CHEMICALS
Chemicals Abbreviation Concentration or purity Supplier
Acetic acid HAc 100% Merck1
Acetonitrile ACN Ultra LC-MS grade Actu-All2
Ammonium hydroxide NH4OH 25% VWR3
Ammonium formate Amm.form >97% Sigma-Aldrich4
Formic acid FA 98-100% Merck1
n-Hexane HPLC grade Actu-All2
Methanol MeOH Ultra LC-MS grade Actu-All2
Periodic acid >99% Sigma-Aldrich4
Water H2O Ultra LC-MS grade Actu-All2
1
Merck, Darmstadt, Germany
2 Actu-All, Oss, The Netherlands 3
VWR International, Amsterdam, The Netherlands
4 Sigma-Aldrich, Zwijndrecht, The Netherlands
TABLE 3: HYDROPHILIC STANDARDS
Chemicals Abbreviation Concentration or purity Supplier L-2-Amino-3-methylaminopropionic acid BMAA >97% Sigma-Aldrich1
Anatoxin ATX 4.96 ± 0.18 μg mL-1 NRC2 Cylindrospermopsin CYN 12.6 ± 0.8 μg mL-1 NRC2 Decarbamoylgonyautoxin 2&3 dcGTX2 40.9 ± 1.8 μg mL -1 NRC2 dcGTX3 9.2 ± 0.3 μg mL-1 Decarbamoylsaxitoxin dcSTX 16.7 ± 0.5 μg mL-1 NRC2 Decarbamoylneosaxitoxin dcNEO 8.0 ± 0.3 μg mL-1 NRC2 L-2,4-Diaminobutyric acid DBA >95% Sigma-Aldrich1
Domoic acid DA >90% Sigma-Aldrich1
Gonyautoxin 1&4 GTX1 24.8 ± 1.3 μg mL -1 NRC2 GTX4 8.1 ± 0.7 μg mL-1 Gonyautoxin 2&3 GTX2 45.2 ± 2.3 μg mL -1 NRC2 GTX3 17.2 ± 0.9 μg mL-1 Gonyautoxin 5 GTX5 24.7 ± 1.1 μg mL-1 NRC2 Neosaxitoxin NEO 20.7 ± 1.1 μg mL-1 NRC2 Saxitoxin STX 19.8 ± 0.4 μg mL-1 NRC2 N-Sulfocarbamoylgonyautoxin-2&3 C1 53.9 ± 1.8 μg mL -1 NRC2 C2 16.1 ± 1.3 μg mL-1 Tetrodotoxin TTX 96% Latoxan3
Tetrodotoxin & 4,9-anhydro TTX TTX 25.6 ± 1.8 µg mL
-1
CIFGA4 anhTTX 3.0 ± 0.2 µg mL-1
1 Sigma-Aldrich, Zwijndrecht, The Netherlands 2
National Research Council,Measurement science and standards, Halifax, Canada
3 Latoxan, Valence, France 4 CIFGA, Lugo, Spain
14
TABLE 4: LIPOPHILIC STANDARDS
Chemicals Abbreviation Concentration or purity Supplier Azaspiracid-1 AZA1 1.24 ± 0.07 µg mL-1 NRC1
Azaspiracid-2 AZA2 1.28 ± 0.05 µg mL-1 NRC1 Azaspiracid-3 AZA3 1.04 ± 0.04 µg mL-1 NRC1 Azaspiracid-4 AZA4 1.19 ± 0.07 μg mL-1 CIFGA2 Azaspiracid-5 AZA5 1.20 ± 0.07 μg mL-1 CIFGA2
Brevetoxin 2 PbTx 2 95% Latoxan3
Brevetoxin 3 PbTx 3 95% Latoxan3
Brevetoxin 9 PbTx 9 95% Latoxan3
13-Desmethyl spirolide C SPX1 7.0 ± 0.4 µg mL-1 NRC1 13,19-Didesmethyl spirolide C 13,19-didesMeSPXC 10.24 ± 0.98 μg mL-1 CIFGA2
Dinophysistoxin-1 DTX1 15.1 ± 1.1 µg mL-1 NRC1
Dinophysistoxin-2 DTX2 7.8 ± 0.4 µg mL-1 NRC1
Gymnodimine GYM 5.0 ± 0.2 µg mL-1 NRC1
Homoyessotoxin hYTX 5.8 ± 0.3 µg mL-1 NRC1 20-Methyl spirolide G 20MeSPXG 7.01 ± 0.61 µg mL-1 CIFGA2
Microcystin-HilR MC-HilR >95% Enzo Life Sciences4
Microcystin-HtyR MC-HtyR >95% Enzo Life Sciences4
Microcystin-LA MC-LA >95% Enzo Life Sciences4 Microcystin-LF MC-LF >95% Enzo Life Sciences4 Microcystin-LR MC-LR >95% Enzo Life Sciences4
[D-Asp3]Microcystin-LR Asp MC-LR >95% Enzo Life Sciences4
Microcystin-LW MC-LW >95% Enzo Life Sciences4 Microcystin-LY MC-LY >95% Enzo Life Sciences4 Microcystin-RR MC-RR >95% Enzo Life Sciences4 Microcustin-YR MC-YR >95% Enzo Life Sciences4
Nodularin NOD >95% Enzo Life Sciences4
Okadaic acid OA 13.7 ± 0.6 µg mL-1 NRC1
Okadaic acid C8-diol ester OA C8-diol ester >90% Enzo Life Sciences4
Okadaic acid methyl ester OA methyl ester >90% Enzo Life Sciences4 Pacific ciguatoxin 1 pCTX1 No certified concentration University of Queensland5
Pacific ciguatoxin 2 pCTX2 No certified concentration University of Queensland5
Pacific ciguataxin 3 pCTX3 No certified concentration University of Queensland5 7-O-Palmitoylokadaic acid 16:0 OA ester, DTX3 90-94% MP Biomedicals6
Palytoxin PlTX >90% Wako7
Pectenotoxin-2 PTX2 8.6 ± 0.3 µg mL-1 NRC1
Pinnatoxin E PnTX E No certified concentration Cawthron Institute8
Pinnatoxin F PnTX F No certified concentration Cawthron Institute8 Pinnatoxin G PnTX G No certified concentration Cawthron Institute8
Yessotoxin YTX 5.6 ± 0.2 µg mL-1 NRC1
1
National Research Council,Measurement science and standards, Halifax, Canada
2
CIFGA, Lugo, Spain
3
Latoxan, Valence, France
4
Enzo Life Sciences, Antwerp, Belgium
5
Professor Lewis, Institute for molecular Bioscience, The University of Queensland, Australia
6
MP Biomedicals, Santa Ana, United states
7 Wako, Osaka, Japan
15
2.2
P
REPARATION OF STANDARD SOLUTIONS
For method development 5 mixtures were prepared. Standard mixture 1 contained all hydrophilic phycotoxins except PSP toxins, at a concentration of 1 µg mL-1 in water: TTX, anhTTX (117 ng mL-1), BMAA, DBA, DA, ATX and CYN. Standard mixture 2 contained all PSP toxins at a concentration of 1 µg mL-1 in water containing 0.03 M acetic acid: STX, NEO, dcSTX, dcNEO, GTX1&4 (GTX4 325 ng mL-1), GTX2&3 (GTX3 380 ng mL-1), GTX5, dcGTX2&3 (dcGTX3 224 ng mL-1) and C1&2 (C2 299 ng mL-1). Standard mixture 3 contained all microcystins at a concentration of 1 µg mL-1 in methanol/water (80:20 v/v): LA, LF, LR, LW, LY, RR, MC-WR, MC-HilR, MC-HtyR, MC-YR, Asp MC-LD and NOD. Standard mixture 4 contained lipophilic toxins at a concentration of 100 ng mL-1 in methanol: OA, DTX1, DTX2, YTX, hYTX, PTX2, SPX1, GYM, 13,19-didesMeSPXC, 20MeSPXG, 16:0 OA ester, OA C8-diol ester, AZA1, AZA2, AZA3 and AZA4. Standard mixture 5 contained all other lipophilic phycotoxins at a concentration of 100 ng mL-1 in methanol: PnTX E, PnTX F, PnTX G, pCTX1 (50 ng mL-1), pCTX2 (50 ng mL-1), pCTX3 (50 ng mL-1), PbTx2, PbTx3, PbTx9, PlTX, AZA5, OA methyl ester and DA. For the validation some phycotoxins were excluded due to lack of standards (pCTX1, 2 and 3), poor sensitivity (PlTX) or poor peak shapes (DBA and BMAA) during method development. For the validation the standards of mixtures 4 and 5 were combined in a single mixture.
2.3
P
REPARATION OF EXTRACTS
T
ISSUE SAMPLESThe extraction method of phycotoxins from tissue samples like fish and shellfish was developed and optimised. Multiple extraction solvents and volumes were tested with naturally contaminated samples. Additionally some different extraction methods were tested. Recoveries were obtained with spiked samples and compared to existing confirmatory methods for PSP toxins, DSP toxins, AZAs, and MCs.
The optimised extraction procedure was as follows: 1.0 ± 0.05 g tissue homogenate was weighed and extracted with 4 mL methanol. The sample was vortex mixed for one minute using a multipulse vortex. The sample was centrifuged at 2,000 x g for 5 minutes and the supernatant was decanted from the pellet to a graduated tube. 5 mL of H2O/ACN/Amm.form/FA (55:45 v/v, 2 mM, 0.5 mM) was added to the pellet. The sample was again vortex mixed for one minute using a multipulse vortex. The sample was centrifuged at 2,000 x g for 5 minutes and the supernatant was combined with the previous obtained methanol extract. The tube was filled to 10 mL with acetonitrile. To avoid loss of compounds two different filters were used suitable for lipophilic and hydrophilic phycotoxins. For the analysis of lipophilic phycotoxins an aliquot of the extract was filtered with a 0.2 µm HT Tuffryn filter (Sigma-Aldrich, Zwijndrecht, The Netherlands). For analysis of hydrophilic phycotoxins an aliquot of the extract was filtered with a 0.45 µm PVDF filter (Sigma-Aldrich, Zwijndrecht, The Netherlands). The filtered extracts were transferred into a glass vial and used for analysis with LC-hrMS.
W
ATER SAMPLESFor the analysis of the phycotoxins in water a clean-up procedure was developed and further optimised. Two clean-up methods are needed, one specific for lipophilic phycotoxins and one for hydrophilic phycotoxins. For both phycotoxin groups solid phase extraction (SPE) is used. Because water samples might contain algal cells which can hold phycotoxins, a method to disrupt whole algal cells was optimised. It was assumed that the disruption method with the maximal yield of lipophilic phycotoxins is also suitable for hydrophilic phycotoxins. For the clean-up procedure of lipophilic phycotoxins, the SPE cartridge was selected which gave the best results for regulated lipophilic phycotoxins in shellfish extracts [65]. The washing step of the SPE procedure was further optimised for a wider range of compounds in water samples. For hydrophilic phycotoxins SPE cartridges
16 were tested in order to reduce matrix effects and to concentrate the hydrophilic toxins. Another option tested to concentrate the phycotoxins was by evaporation. Recoveries were obtained with spiked samples and compared to a standard solution for PSP toxins, regulated lipophilic phycotoxins, some CIs and MCs.
The optimised clean-up procedures were as follows: for the clean-up of water containing lipophilic phycotoxins a 30 mg Strata-X polymeric reversed phase cartridge (Phenomenex, Utrecht, The Netherlands) was used. The cartridge was activated and conditioned with 1 mL methanol followed by 1 mL water. A water sample of 1 mL was loaded onto the cartridge and the cartridge was washed with 1 mL water. Subsequently, the lipophilic phycotoxins were eluted with 1 mL methanol. The eluent was transferred into a glass vial and used for analysis with LC-hrMS. To extract hydrophilic phycotoxins from water a 500 mg Chromabond HILIC cartridge (Macherey-Nagel, Düren, Germany) was used. The cartridge was activated and conditioned with 1 mL water followed by 6 mL acetonitrile. A water sample of 1 mL diluted with 9 mL acetonitrile was loaded onto the cartridge and subsequently washed with 2 mL acetonitrile. The hydrophilic phycotoxins were eluted with 2 mL water. The eluent was diluted with 2 mL acetonitrile and transferred into a glass vial for analysis with LC-hrMS.
F
OOD SUPPLEMENTSThe extraction and clean-up procedure for food supplements was based on the methods for tissue and water. Pills were grinded or capsules were removed on forehand. 1.0 ± 0.05 g of food supplement was weighed and extracted with 4 mL methanol. The sample was vortex mixed for one minute using a multipulse vortex. Subsequently the sample was ultrasonic disrupted for 1 minute at 11 W to disrupt possible whole algal cells present. The sample was centrifuged at 2,000 x g for 5 minutes and the supernatant was decanted from the pellet to a graduated tube. 5 mL of H2O/ACN/Amm.form/FA (55:45 v/v, 2 mM, 0.5 mM) was added to the pellet. The sample was vortex mixed for one minute using a multipulse vortex. The sample was centrifuged at 2,000 x g for 5 minutes and the supernatant was combined with the methanol extract. The tube was filled to 10 mL with acetonitrile. The extract was diluted with 67.5 mL water in order to obtain a 10% organic strength prior to the SPE procedure for lipophilic phycotoxins. A 60 mg Strata-X polymeric reversed phase cartridge was activated and conditioned with 3 mL methanol and 3 mL water. The diluted extract was loaded onto the cartridge and was washed with 3 mL water. Respectively, the lipophilic phycotoxins were eluted with 2 mL methanol and the eluate was transferred to a sample vial for analysis with LC-hrMS. The clean-up method for hydrophilic phycotoxins in food supplements is still in development. When it is developed all quantities of the extraction procedure can be doubled and then the extract can be split for both clean-up methods.
E
XTRACTION OF PALYTOXIN-
GROUP TOXINS FOR CONFIRMATIONSelwood et al. developed a method for fish and shellfish to analyse PlTXs like palytoxins, ovatoxins and ostreocins using LC-MS after micro-scale oxidation [58]. The oxidative fragmentation reaction causes the PlTXs to fragment to an amide aldehyde specific oxidation product for each compound and a common amino aldehyde as shown in Figure 1. The method was further optimised. Multiple extraction solvents were tested; LC and the MS conditions were optimised. Subsequently repeatability, extraction efficiency, matrix effects, limit of detection (LOD), limit of quantitation (LOQ) and linearity were determined in shellfish extracts for the oxidation products: amide aldehyde of palytoxin and the common amino aldehyde.
17
FIGURE 1: OXIDATIVE FRAGMENT PRODUCTS OF PALYTOXIN
The optimised extraction procedure was as follows: 1.0 ± 0.05 g tissue homogenate was weighed and extracted with 3 mL methanol. The sample was vortex mixed for one minute using a multipulse vortex. The sample was centrifuged at 2,000 x g for 5 minutes and the supernatant was decanted from the pellet to a graduated tube. The extraction was repeated twice. After the third extraction 9 mL water was added to the 9 mL supernatant to dilute the extract before solid phase extraction was applied. A 60 mg Strata-X polymeric reversed phase cartridge was conditioned with 3 mL methanol followed by 3 mL water. Subsequently the entire diluted sample extract was loaded onto the cartridge. The cartridge was washed with 2 mL methanol/water (40:60 v/v) followed by 2 mL water and then the sample was oxidized with 2 mL of 50 mM periodic acid. The cartridge was washed for a second time with 2 mL water. Thereafter the oxidation products were eluted with 3 mL MeOH/H2O/HAc (60:40:0.1 v/v). The eluent was transferred into a glass vial and used for analysis with LC-MS/MS.
2.4
I
NSTRUMENTATION
S
CREENING OF PHYCOTOXINS BYLC-
HRMS
For the screening of phycotoxins a Thermo Scientific UltiMate 3000 LC-system (Thermo Fisher Scientific, Waltham, USA) coupled to a Thermo Scientific Q Exactive focus hybrid quadrupole-orbitrap mass spectrometer was used. Hydrophilic phycotoxins are not well retained on a reversed phase column and the lipophilic phycotoxins are not well retained on the HILIC column. Therefore it was decided to develop two different LC methods based on reversed phase and HILIC conditions. The mobile phase compositions were kept identical for both LC methods. Mobile phase A consisted of water and mobile phase B consisted of acetonitrile/water (9:1 v/v), both containing 2 mM ammonium formate and 0.5 mM formic acid. Chromatographic separation of
H2N O OH OH OH O OH OH HO OH OH OH OH O OH OH OH OH HO O OH OH HO OH OH OH O OH OH OH HO OH OH OH OH HO OH O OH OH HO N H N H HO OH OH OH OH O O OH OH HO O O O O H H
common amino aldehyde
18 lipophilic phycotoxins was achieved on a reversed phase ACQUITY BEH C18 1.7 µm, 100 · 2.1 mm UPLC column (Waters, Milford, MA, USA). The column temperature was set at 35°C and the total run time was 28 minutes. The gradient elution with a flow of 0.3 mL min-1 was as follows: 0.1 minute at 10% mobile phase B, then linearly increased to 100% mobile phase B in 12.9 minutes and kept at 100% mobile phase B for 12 minutes. Subsequently the gradient went back to 10% mobile phase B in 0.1 minute and kept at 10% mobile phase B for 2.9 minutes to equilibrate the column for the next run. Chromatographic separation of hydrophilic phycotoxins was achieved on a TSKgel Amide-80 2 µm, 150 · 3 mm HPLC column (Tosoh Bioscience, Tokyo, Japan). The column temperature was set at 35°C and the total run time was 20 minutes. The gradient elution with a flow of 0.5 mL min-1 was as follows: 0.1 minute at 90% mobile phase B, then linearly decreased to 45% mobile phase B in 13.9 minutes and subsequently linearly decreased to 20% mobile phase B in 0.1 minute and kept at 20% mobile phase B for 1.9 minutes. Subsequently the gradient went back to 90% mobile phase B in 0.1 minute and kept at 90% mobile phase B for 3.9 minutes to equilibrate the column. For both chromatographic methods the injection volume was set at 10 µL.
In order to detect the phycotoxins, electrospray ionisation (ESI) in both positive and negative mode was used. The positive and negative ESI signals were acquired in two separate runs. The spray voltage in positive ionisation mode was set at 3.5 kV and in negative ionisation mode at -2.5 kV. The capillary temperature was set at 260°C. A full MS scan event of 100 to 1500 m/z with a resolution of 70,000 full width at half maximum (FWHM) was acquired. In order to obtain additional information on the phycotoxins, fragmentation spectra were also acquired. The so called MS2 scans were obtained by selecting all ions in respectively m/z mass range windows of 100 to 500, 500 to 1000 and 1000 to 1500. As collision gas nitrogen was used. The normalized collision energy (NCE) was set at 40 during fragmentation of all ion mass ranges, except for 100 to 500 m/z in negative ionisation mode where the NCE was set at 30. Then after fragmentation the ions were scanned respectively from 50 to 500, 50 to 1000 and 50 to 1500 m/z with a resolution set at 17,500 FWHM. The automatic gain control representing the maximum capacity of the C-trap was set at a maximum of 106 ions or a maximum injection time of 200 ms for both the full scan and MS2 scans were allowed.
To process the data Tracefinder (Thermo Fisher Scientific, Waltham, USA) was used. A database containing over 800 phycotoxins was constructed from literature. A part of the database is shown in Appendix 2 for illustration. Relevant parts of the database for data processing were transferred to a Tracefinder compound database and data was searched. There was a mass error allowed of 5 parts per million (ppm) for the m/z of the precursor ions and for regulated phycotoxins at least one fragment ion should be present within a 5 ppm mass error. Furthermore, if the retention time of a compound was known it should be within a ± 0.2 minutes retention window.
C
ONFIRMATION OF PALYTOXIN-
GROUP TOXINS BYLC-MS/MS
For PlTXs a Waters Acquity UPLC system coupled to a Waters Xevo TQ-S tandem mass spectrometer (Waters, Milford, USA) was used. Chromatographic separation was achieved by a Waters Acquity BEH C18 100 · 2.1 mm, 1.7 µm UPLC column at 80°C, with a gradient of 5 minutes. Mobile phase A consisted of water and mobile phase B consisted of acetonitrile/water (9:1 v/v), both containing 6.7 mM ammonium hydroxide. The gradient with a flow of 0.6 mL min-1 was kept the first 0.5 minutes at 0% mobile phase B, and then it linearly increased to 90% mobile phase B in 3 minutes and kept at 90% mobile phase B for 0.5 minutes. Thereafter the gradient went back to 0% mobile phase B in 0.1 minutes and kept for 0.9 minutes at 0% mobile phase B to equilibrate the column for the next run. The injection volume was set at 10 µL.
The MS system was operated in positive electrospray mode and data was recorded in MRM mode using two transitions per PlTX oxidative fragmentation product. Table 5 shows the MS/MS conditions used to acquire the
19 MRM data. Furthermore, a capillary voltage of 3 kV, source temperature of 150 °C, desolvation temperature of 500 °C and desolvation gas flow of 800 L h-1 was set.
TABLE 5: MS/MS CONDITIONS FOR THE CONFIRMATION OF PALYTOXIN-GROUP TOXINS
Compound Precursor ion(m/z) Product ion (m/z) Cone (V) Collision energy (eV)
Common amino aldehyde 300.2 107.0 30 25
300.2 151.0 30 25 Neo-palytoxin amide aldehyde 325.2 76.0 30 20 325.2 307.0 30 20 Ostreocin D amide aldehyde 329.2 76.0 30 13 329.2 149.0 30 13 Palytoxin, Tahitian palytoxin, Ovatoxin-a, 343.2 76.0 30 13 42-OH palytoxin amide aldehyde 343.2 123.0 30 13 Homopalytoxin amide aldehyde 357.2 76.0 30 20 357.2 339.0 30 20 Bishomopalytoxin amide aldehyde 371.2 76.0 30 20 371.2 353.0 30 20
2.5
V
ALIDATION
The following phycotoxins were validated: OA, DTX1, DTX2, YTX, hYTX, PTX2, SPX1, GYM, 13,19 didesm SPX C, 20 meth SPX G, OA methyl ester, 16:0 OA ester, OA C8 diol ester, AZA1, AZA2, AZA3, AZA4, AZA5, PnTX E, PnTX F, PnTX G, CYN, DA, MC-LA, MC-LF, MC-LR, MC-LW, MC-LY, MC-RR, MC-WR, MC-HilR, MC-HtyR, MC-YR, Asp MC-LR, NOD, STX, dcSTX, NEO, dcNEO, GTX1&4, GTX2&3, GTX5, dcGTX2&3, C1&2, TTX and ATX. The validation of the screening method was based on the estimated screening detection limit (SDL). The SDL of the qualitative screening method is the lowest level at which an analyte has been detected in at least 95% of the samples. 20 blank tissue samples, including 5 mussel-, 5 oyster-, 5 cockle-, 5 ensis- and 5 fish homogenates, were spiked after extraction with 600 µg kg-1 hydrophilic phycotoxins, 150 µg kg-1 microcystins and 80 µg kg-1 of all other lipophilic phycotoxins included in the validation. Five blanks, one of each matrix, were included to determine false positives. 20 blank water samples, including 6 sea water, 6 brackish water, 6 fresh water and 2 tap water were spiked before clean-up with 120 µg L-1 hydrophilic phycotoxins, 10 µg L-1 MCs and 5 µg L-1 of all other lipophilic phycotoxins. DA was included in both clean-ups for hydrophilic phycotoxins and lipophilic phycotoxins. Furthermore 4 blanks were included to determine false positives. 20 blank solid food supplements were spiked before extraction with 30 µg kg-1 MCs and 15 µg kg-1 of all other lipophilic phycotoxins. 20 blank liquid food supplements (mostly oils) were spiked before extraction with 50 µg kg-1 MCs and 15 µg kg-1 of all other lipophilic phycotoxins. 9 blank food supplements were included to determine false positives. For each procedure also an empty tube was included to determine if there were any interfering contaminants from the procedures itself. Furthermore, the confirmation and quantitation of regulated lipophilic phycotoxins and some CIs, PSP toxins and DA in shellfish were validated: OA, DTX1, DTX2, YTX, hYTX, SPX1, GYM, AZA1, AZA2, AZA3, PnTX G, DA, STX, dcSTX, NEO, dcNEO, GTX1&4, GTX2&3, GTX5 and dcGTX2&3. 5 blank mussel homogenates were spiked after extraction at 0.5 and 1 times the permitted level (PL) according to the European Union. One blank mussel homogenate was spiked before extraction at 0.5 PL and reference materials were included to determine the recovery. Matrix matched standards were spiked at 5 (ASP and PSP) or 6 (lipophilic phycotoxins) levels. Concentrations are shown in Table 6.
20
TABLE 6: SPIKE LEVELS OF ASP, PSP AND DSP FOR QUANTITATIVE VALIDATION IN SHELLFISH [30-34]
Compound Toxin group 0.5 PL
(ug kg-1)
1 PL (ug kg-1)
Matrix matched standards (ug kg-1)
DA ASP 10,000 20,000 0, 5, 10, 20, 50 mg kg-1 STX, dcSTX, NEO, dcNEO, GTX1&41, GTX2&31, GTX5, dcGTX2&31 PSP 400 800 0, 400, 600, 800, 1200 OA, DTX1, DTX2 DSP 80 160 0, 20, 40, 80, 160, 240 AZA1, AZA2, AZA3 80 160 0, 20, 40, 80, 160, 240 YTX, hYTX2 250 500 0, 62.5, 125, 250, 500, 750
SPX13 200 400 0, 50, 100, 200, 400, 600
GYM3 100 200 0, 25, 50, 100, 200, 300
PnTX G3 25 50 0, 6.25, 12.5, 25, 50, 75
1
Concentration of the highest isomer present.
2 The permitted level of YTX and hYTX is 3750 µg kg-1, values are target values. 3
21
3. R
ESULTS AND DISCUSSION
3.1 O
PTIMIZATION OF EXTRACTION PROCEDURES
For development of extraction and clean-up procedures measurements were carried out with available RIKILT standards operation procedures (SOPs). All methods were applied on LC-MS/MS; SOPs are described in Appendix 3. Most recoveries obtained during method development were established by comparing results with results from SOPs or by a mutual comparison between tested procedures. These were comparisons of recoveries of the extraction and/or clean-up procedure including the influence of matrix effects in the purified extracts. Therefore, unless stated otherwise, recoveries reported are apparent recoveries. SOP results were in all cases considered as 100% recovery.
T
ISSUE SAMPLESTo develop an extraction method for both lipophilic and hydrophilic phycotoxins in tissue samples several extraction methods, volumes and duration of the extraction were tested. Natural contaminated samples were used containing OA, DTX2, AZAs, SPX1 and PSP toxins.
To test different extraction methods extractions were carried out with 4 mL methanol, followed by 4 mL H2O/ACN/Amm.form/FA (55:45 v/v, 2 mM, 0.5 mM) or with 4 mL H2O/ACN/Amm.form/FA (55:45 v/v, 2 mM, 0.5 mM), followed by 4 mL methanol. Methanol was used to extract lipophilic phycotoxins and H2O/ACN/Amm.form/FA was used to extract hydrophilic phycotoxins. After adding each extraction solvent samples were vortex mixed for 1 minute, placed in an ultrasonic bath for 5 minutes, ultrasonic disrupted for 30 seconds at 11 W or heated for 5 minutes at 70 °C, followed by centrifugation and decanting of the extract. To obtain a recovery results were compared to the SOP for lipophilic phycotoxins, which is a triplicate extraction with 3 mL methanol, followed by vortex mixing for 1 minute, centrifugation and decanting of the extract after each time adding methanol. The complete SOP extraction procedure is describes in Appendix 3. All extracts were complemented to 10 mL with methanol and filtered. Recoveries (average of n=2) are shown in Table 7.
TABLE 7: AVERAGE RECOVERIES (%) OF LIPOPHILIC PHYCOTOXINS WITH VARIOUS EXTRACTION METHODS
Solvents* Extraction OA DTX2 DTX3
(OA 16:0) AZA1 AZA2 AZA3 SPX1 Average MeOH, MP Vortex 95.2 118.5 89.1 87.8 90.2 91.2 148.6 102.9
100.1 MeOH, MP Ultrasonic bath 85.1 123.6 73.1 78.7 84.4 93.0 130.4 95.5 MeOH, MP Ultrasonic disruptor 89.7 116.4 79.9 83.5 90.1 94.5 136.8 98.7 MeOH, MP Heat 70 °C 76.5 112.3 54.6 89.5 90.4 171.4 128.2 103.3 MP, MeOH Vortex 87.6 121.0 49.9 84.5 81.3 98.6 123.4 92.3
93.0 MP, MeOH Ultrasonic bath 76.1 112.3 38.2 87.8 78.8 131.0 99.6 89.1 MP, MeOH Ultrasonic disruptor 99.2 115.4 52.9 81.3 79.6 119.7 151.0 99.9 MP, MeOH Heat 70 °C 74.2 101.4 37.2 86.9 78.5 157.7 98.1 90.6 *MP = Mobile phase A/B (50:50 v/v), which is H2O/ACN/Amm.form/FA (55:45 v/v, 2 mM, 0.5 mM)
All extraction methods showed good recoveries for the lipophilic phycotoxins present in the samples except for DTX3 and AZA3. On average the recovery was higher and the RSD was lower when starting the extraction with methanol. To obtain a higher recovery for DTX3 a multiple extractions with methanol are needed due to the fatty acid ester chains in the molecules. The recovery of AZA3 in combination with heat was significantly higher because AZA17 present in the sample was converted to AZA3 [66]. Heating was avoided in order to prevent conversion within other samples. For further experiments there was chosen to continue with a combination of vortex mixing and ultrasonic disruption.
22 To optimize the extraction procedure for lipophilic phycotoxins in tissue the duration of vortex mixing and ultrasonic disruption was varied. 4 mL methanol was added to the samples, the sample was vortex mixed for 1 or 5 minutes, followed by centrifugation and decanting of the extract. Subsequently, 4 mL H2O/ACN/Amm.form/FA (55:45 v/v, 2 mM, 0.5 mM) was added and ultrasonic disrupted for 30 seconds, 1 minute or 5 minutes at 11 W, followed by centrifugation and decanting of the extract. The results were compared to the results of the SOP for lipophilic phycotoxins to obtain a recovery. Recoveries (average of n=2) are shown in Table 8.
TABLE 8: AVERAGE RECOVERIES (%) OF LIPOPHILIC PHYCOTOXINS WITH VARIOUS DURATION OF EXTRACTION
Vortex Ultrasonic
disruptor OA DTX2
DTX3
(OA 16:0) AZA1 AZA2 AZA3 SPX1 Average 1 minute 0.5 minute 101.2 94.0 44.4 112.3 104.9 92.3 87.4 90.9 91.2 1 minute 1 minute 121.1 127.9 48.5 89.6 85.0 87.2 85.6 92.1 1 minute 5 minutes 108.0 119.2 43.7 96.3 95.3 91.7 80.2 90.6 5 minutes 0.5 minute 94.7 99.3 27.6 104.8 103.1 94.2 79.4 86.2 84.5 5 minutes 1 minute 80.0 67.0 14.2 102.8 103.3 93.1 69.4 75.7 5 minutes 5 minutes 103.8 109.7 37.7 105.2 106.6 97.2 80.4 91.5
A longer time of vortex mixing or ultrasonic disruption did not have an effect on the recovery. Because a longer extraction time did not have a positive effect on the recovery there was continued with 1 minute of vortex mixing and 1 minute of ultrasonic disrupting. After the lipophilic procedure, extraction optimization for the hydrophilic phycotoxins was optimised. Different extraction volumes were tested. Natural contaminated samples containing PSP toxins were used. 4 mL methanol was added to the samples, the sample was ultrasonic disrupted for 1 minute at 11 W, followed by centrifugation and decanting of the extract. Subsequently, 4 or 5 mL H2O/ACN/Amm.form/FA (55:45 v/v, 2 mM, 0.5 mM) was added and vortex mixed for 1 minute, followed by centrifugation and decanting of the extract. To obtain a recovery, results were compared to the results of a triplicate extraction with 3 mL H2O/ACN/Amm.form/FA (55:45 v/v, 2 mM, 0.5 mM), followed by vortex mixing for 1 minute, centrifugation and decanting of the extract after each time adding extraction solvent. All extracts were complemented to 10 mL with acetonitrile and filtered. Recoveries (average of n=2) are shown in Table 9.
TABLE 9: AVERAGE RECOVERIES (%) OF HYDROPHILIC PHYCOTOXINS WITH VARIOUS VOLUMES OF EXTRACTION SOLVENT
MeOH MP* dcSTX NEO dcNEO GTX3 GTX4 GTX5 dcGTX3 C2 C1 GTX6 Average 4 mL 4 mL 99.1 56.3 99.6 82.7 96.8 94.2 94.3 91.9 55.8 94.9 86.6 4mL 5 mL 94.0 87.8 99.1 83.7 99.5 94.8 89.0 90.2 55.9 94.4 88.8 *MP = Mobile phase A/B (50:50 v/v), which is H2O/ACN/Amm.form/FA (55:45 v/v, 2 mM, 0.5 mM)
On average the recovery was slightly higher when 5 mL H2O/ACN/Amm.form/FA (55:45 v/v, 2 mM, 0.5 mM) was used during the second extraction step. Especially the recovery for NEO is improved from 56% to 88%. To verify the recovery of other phycotoxins blank shellfish samples were spiked with standards and to further improve the efficiency of the method, ultrasonic disrupting was replaced by vortex mixing. 4 mL methanol was added to the samples, the sample was ultrasonic disrupted or vortex mixed for 1 minute, followed by centrifugation and decanting of the extract. Subsequently, 5 mL H2O/ACN/Amm.form/FA (55:45 v/v, 2 mM, 0.5 mM) was added and vortex mixed for 1 minute, followed by centrifugation and decanting of the extract. To obtain a recovery the results were compared to the results of samples extracted with SOPs for lipophilic toxins, PSP toxins and MCs. Extraction procedures of the SOPs are described in Appendix 3. Recoveries (average of n=2) are shown in Table 10.
23
TABLE 10: AVERAGE RECOVERIES (%) OF SPIKED STANDARDS WITH AND WITHOUT ULTRASONIC DISRUPTION
First extraction OA DTX1 DTX2 AZA1 AZA2 AZA3 AZA4 YTX hYTX SPX1 GYM Average Ultrasonic disruptor 119.5 123.9 120.3 91.7 90.8 98.1 91.2 96.9 103.9 100.9 101.5
Vortex 131.5 120.9 128.0 93.3 91.5 93.9 87.9 92.5 99.5 117.9 106.3
STX dcSTX NEO dcNEO GTX3 GTX4 GTX5 dcGTX3 TTX DA ATX CYN Ultrasonic disruptor 79.9 87.8 78.9 111.0 67.2 157.6 55.1 106.8 111.1 101.9 200.6 109.1 Vortex 81.7 88.8 78.3 97.6 73.5 131.5 69.9 101.3 106.8 94.6 204.4 118.1 MC- WR HtyR YR LW HilR LY LR LF aspLR LA NOD RR
Ultrasonic disruptor 149.6 110.7 105.7 157.9 119.9 158.0 128.7 97.9 120.2 140.2 88.7 158.1 112.6 Vortex 167.8 129.5 119.7 146.1 129.8 162.2 133.9 88.4 132.0 141.4 104.7 152.9 114.8
All recoveries are above 80% except for some PSP toxins. On average the recovery for extraction with only vortex mixing is higher than an extraction with ultrasonic disruption. The final procedure is described in the materials and methods chapter.
W
ATER SAMPLESWater samples can contain algal cells. Assumed was that phycotoxins can be present in algal cells and that those cells might not break open during a clean-up procedure. When cells do not break open during clean-up, the phycotoxins present in the whole cells are not measured and results give unreliable information about the toxicity of the sample. To make sure cells were disrupted different disrupting methods were tested with a stain of Alexandrium Ostenfeldii from the Ouwerkerkse kreek in The Netherlands producing SPX1 and GYM.
Parts of the sample were ultrasonic disrupted for 1 minute at 11W, placed in an ultrasonic bath for 5 minutes, frozen, 20 seconds grinded twice with a Precellys (VWR, Amsterdam, The Netherlands) at 6500 rpm or used without any treatment. Then all whole cells were removed by using a 0.2 µm HT Tuffryn filter (Sigma-Aldrich, Zwijndrecht, The Netherlands). Furthermore, two filters used for samples without treatment were washed with 5 mL water or 1 mL methanol to determine any osmotic effects or effects of organic solvents on the algae. Subsequently, 1 mL of the filtered samples was cleaned-up with solid phase extraction (SPE). Also a sample without any treatment or filtration was cleaned-up. Methanol was diluted with 4 mL water before SPE to ensure retention of SPX1 and GYM on the SPE cartridge. A 30 mg Strata-X polymeric reversed phase cartridge (Phenomenex, Utrecht, The Netherlands) was activated and conditioned with 1 mL methanol followed by 1 mL water. A water sample of 1 mL was loaded onto the cartridge and the cartridge was washed with 1 mL water. Subsequently, the lipophilic phycotoxins were eluted with 1 mL methanol. Because the concentration of SPX1 and GYM in the water sample was unknown, the result with the highest response was set at a 100% recovery, other result were compared to this. The recoveries (average of n=2) are shown in Table 11.
TABLE 11: AVERAGE RECOVERIES (%) AFTER ALGAL DISRUPTION
Sample treatment SPX1 GYM Filter – SPE 14.4 12.8 Ultrasonic disruptor – filter – SPE 68.8 78.8 Ultrasonic bath – filter – SPE 15.6 14.8 Frozen – filter – SPE 58.9 66.7 Precellys – filter – SPE 53.7 67.0 Filter – water wash – SPE 12.6 14.8 Filter – MeOH wash – SPE 40.2 46.5
SPE 100.0 100.0
With filtration only phycotoxins present in the water were measured. It appears that only a small amount (14%) of SPX1 and GYM was excreted by the algae into the water. Ultrasonic disruption, freezing, grinding with the
24 Precellys and treatment with methanol disrupted some of the algae and phycotoxins were released. There was no osmotic effect because the concentration after washing with water was not increased compared to only filtration. However, the best results were achieved with use of only the SPE clean-up. The experiment was repeated without the filtration step; nevertheless, there was no increase in recovery when ultrasonic disrupted samples were directly cleaned-up with SPE. This could mean that all toxins are released from the algal cells during the SPE procedure. Most probably the cells were lysed during the addition of 1 mL methanol.
Brackish medium was used to cultivate the strain of Alexandrium Ostenfeldii. Therefore, to test the SPE clean-up method for other lipophilic phycotoxins a blank brackish medium was spiked with lipophilic phycotoxins. Different washing steps were tested to optimize the method. A 30 mg Strata-X polymeric reversed phase cartridge was activated and conditioned with 1 mL methanol followed by 1 mL water. A water sample of 1 mL was loaded onto the cartridge and the cartridge was washed with 1 mL water or water/methanol (80:20 v/v). Subsequently, the lipophilic phycotoxins were eluted with 1 mL methanol. The results (average of n=2) were compared to a standard solution to obtain a recovery. Results are shown in Table 12.
TABLE 12: AVERAGE RECOVERIES (%) OF LIPOPHILIC PHYCOTOXINS IN WATER AFTER SOLID PHASE EXTRACTION
Washing step OA DTX1 DTX2 AZA1 AZA2 AZA3 AZA4 YTX hYTX SPX1 GYM PnTX G Average Water 148.0 86.8 141.1 41.4 33.5 49.7 57.6 120.9 98.6 79.1 85.2 70.1 84.3 20% MeOH 124.0 83.6 125.1 50.6 41.1 61.9 69.7 93.1 89.9 92.3 102.8 84.9 84.9
MC- WR HtyR YR LW HilR LY LR LF aspLR LA NOD RR
Water 29.5 63.8 61.9 10.4 40.2 42.8 54.6 18.7 35.1 47.7 75.8 97.2 48.1 20% MeOH 29.6 40.9 48.6 10.2 32.4 36.1 48.1 16.2 31.8 39.4 76.1 67.8 39.8
Besides the compounds in Table 12, DA was analysed as well but had no recovery. Therefore concluded was that DA was not retained on the SPE cartridge, since DA is also not retained on the reversed phase LC column with the conditions of this developed screening method. Recoveries of the MCs appear to be low (10 to 97%), compared to a standard solution without clean-up, however compared to the results of the SOP for MCs in water (data not shown, procedure in appendix 3) recoveries are 60 to 480% with an average of 196%. Therefore the developed procedure is an improvement of existing SOP. Some of the compounds gave a recovery >100%, all of these compounds were measured in negative ionisation mode. Although matrix effects are generally less in negative ionisation mode, enhancement of these compounds could be due to matrix effects. Because MCs had a better recovery when washed with water, this was included in the final procedure for lipophilic phycotoxins in water samples. The final procedure is as described before in the materials and methods chapter.
For the hydrophilic phycotoxins a different approach was needed, because hydrophilic phycotoxins do not retain well on a reversed phase cartridge and are eluting simultaneously with salts present in the sample. At first there was attempt to evaporate the sample and reconstitute in a high organic solvent. A high concentration of acetonitrile was needed to trap the phycotoxins of interest on the HILIC LC-column. By evaporating the sample in advance and thereafter dilute with a smaller volume acetonitrile the final concentration would be higher as the initial concentration of the sample and there would be no loss of sensitivity during LC-MS analysis. 100 mL of brackish water sample was evaporated to dryness and then reconstituted in 10 mL of solvents ranging from 10 to 25% water with ACN/Amm.form/FA (90 to 75% v/v, 2 mM, 0.5 mM). Due to the high salt content in the sample two immiscible layers were formed. These extracts cannot be used for LC-MS/MS analysis. Subsequently, different volumes of H2O/ACN/Amm.form/FA (55:45 v/v, 2 mM, 0.5 mM) were used to attempt to reconstitute the sample without formation of immiscible layers.
25 However, even after adding 60 mL there were still two layers present and it was decided that a procedure including evaporation was not applicable for water samples with a relative high salt content. Then SPE with a HILIC cartridge was attempted. Blank brackish water samples of 1 mL were spiked with 100 ng/mL hydrophilic phycotoxins and diluted with 3 mL acetonitrile (75%) or 9 mL acetonitrile (90%) to test how much organic strength was needed for the samples to obtain a good retention on the SPE cartridge. A 500 mg Chromabond HILIC cartridge was activated and conditioned with 1 mL water followed by 6 mL acetonitrile. The diluted water sample was loaded onto the cartridge and subsequently washed with 2 mL acetonitrile. The hydrophilic phycotoxins were eluted with 2 mL water. The effluent of the sample, washing solvent and eluent were collected. The effluent of the sample was evaporated and reconstituted in 75% acetonitrile, the washing solvent was diluted with water and the eluent was diluted with acetonitrile. All fractions (n=2) were compared to a standard solution to obtain a recovery. The results (average of n=2) are given in Table 13.
TABLE 13: AVERAGE RECOVERIES (%) OF HYDROPHILIC PHYCOTOXINS WITH CHROMABOND SOLID PHASE EXTRACTION
Organic strength
SPE fraction
STX dcSTX NEO dcNEO GTX3 GTX4 GTX5 dcGTX3 C 1 TTX DA ATX CYN 75% Sample 10.4 10.8 9.7 7.7 15.7 0.8 16.1 15.0 0.1 16.4 0.2 52.0 36.2 Wash 21.3 24.2 38.8 26.0 20.3 6.3 22.4 18.4 0.2 24.7 0.1 26.1 19.7 Elution 51.6 55.6 84.8 67.8 46.0 57.9 56.7 56.1 3.0 42.1 7.6 1.7 18.7 Total 83.3 90.6 133.3 101.4 82.0 65.0 95.2 89.5 3.4 83.1 7.8 79.8 74.6 90% Sample 0.1 0.3 0.1 5.4 0.1 0.0 0.1 0.1 0.1 0.1 0.7 55.5 0.2 Wash 0.1 0.2 0.4 1.9 0.1 0.0 0.1 0.1 0.1 0.1 0.2 22.4 1.0 Elution 74.4 66.7 114.7 101.0 68.3 58.9 77.5 75.3 2.7 78.9 7.2 3.2 69.4
When the sample was diluted to only 75% acetonitrile most hydrophilic phycotoxins were not well retained on the cartridge as most compounds eluted with loading of the sample and during the washing step. However, when the sample was diluted to 90% acetonitrile most hydrophilic phycotoxins were well retained during solid phase extraction, except for ATX which was still eluting during loading of the sample and the washing step. C1 and DA were showing bad recoveries; although there is no clear explanation it might that these phycotoxins precipitated during dilution of the sample together with the salts, were not eluted from the cartridge or were suppressed due to matrix effects during LC-MS analysis. The final procedure is described in the materials and methods chapter.
F
OOD SUPPLEMENTSThere is a large variety of food supplements as shown in Figure 2. Samples can be oils, pills or powders. It can be expected that extraction behaviour for these matrices are difficult and it might be impossible to obtain good recoveries for all supplements that are available. Still the procedure which is a combination of the tissue and water procedure was tested. A true recovery of the extraction and SPE was determined by spiking food supplements before extraction and spiking another extract of the same product after SPE. Results (average of n=2) are shown in Table 14.
26
FIGURE 2: VARIATY OF FOOD SUPPLEMENTS USED FOR RECOVERY EXPERIMENTS
TABLE 14: AVERAGE RECOVERIES (%) OF LIPOPHILIC PHYCOTOXINS IN FOOD SUPPLEMENTS
Sample OA DTX1 DTX2 AZA1 AZA2 AZA3 AZA4 YTX hYTX SPX1 GYM PTX2 Average Oil 79.8 80.1 81.8 59.3 46.1 70.3 89.3 62.5 55.9 93.5 92.3 90.7 75.1 Oil capsules 56.2 67.4 66.2 27.4 21.6 39.2 83.8 13.1 9.3 90.8 84.0 64.4 52.0 Powder 91.3 60.8 79.2 49.1 48.2 53.2 57.4 25.0 26.9 62.2 73.5 53.6 56.7 Tablets 77.5 82.5 79.1 34.8 24.1 46.0 69.4 19.5 17.6 79.0 80.1 80.1 75.5 Average 76.2 72.7 76.6 42.7 35.0 52.2 75.0 30.0 27.4 81.4 82.5 72.2
MC- WR HtyR YR LW HilR LY LR LF aspLR LA NOD RR
Oil 84.0 78.8 74.8 80.7 83.5 84.5 86.4 89.5 82.0 72.8 83.6 76.0 81.4 Oil capsules 66.4 56.0 55.2 19.5 53.7 14.9 46.0 17.5 49.8 10.0 41.4 102.5 44.4 Powder 52.6 57.3 53.3 69.3 62.5 53.5 56.8 56.0 57.9 52.8 67.0 84.0 60.2 Tablets 54.4 62.2 64.9 63.3 64.6 44.7 59.6 50.2 64.7 44.1 63.0 50.3 57.2 Average 64.3 63.6 62.0 58.2 66.1 49.4 62.2 53.3 63.6 44.9 63.8 78.2
As expected a lot of different results are obtained for different matrices but also between the phycotoxins. Lowest recoveries were obtained for YTX and hYTX. On average the recovery is 61%, however the recoveries range from 9 to 102%. Because matrix effects are very different from each other it is difficult to develop a method that suits for all types of food supplements with acceptable recoveries for all compounds. Since the clean-up method worked properly for water samples it was assumed that when changing the SPE clean-up method to remove more matrix also some phycotoxins will disappear. Therefore it is not an option to change or add more clean-up steps. When samples are measured in the future, it is recommended that standard addition is applied to determine recoveries of phycotoxins from individual samples.
3.2 C
HROMATOGRAPHY
The separation of lipophilic and hydrophilic compounds is very different. Lipophilic compounds are retained on reversed phase LC-columns. Hydrophilic compounds are retained with HILIC. To increase the efficiency of the method there was attempt to use one generic mobile phase combination of water and organic solvent with additives for both separations.
It was important that peaks were broad enough to create enough data points, as the resolution of the hrMS is related to the scan speed. When a higher resolution is set the scan time is longer. MS measurements were divided in four scan events. One full scan at a resolution of 70,000 FWHM and three MS2 scans at a resolution of 17,500 FWHM with elevated NCE (30% or 40%) as earlier described in the materials and methods chapter. All these events caused a relatively long total cycle time of 0.85 seconds. Therefore, when a minimum of 10 data points per peak are desired, peaks needs to be at least 9 seconds broad.
27
R
EVERSED PHASE LIQUID CHROMATOGRAPHYFor separation of lipophilic phycotoxins an ACQUITY BEH C18 1.7 µm, 100 · 2.1 mm column was used, the same type as used in the SOPs for lipophilic phycotoxins (Appendix 3). Mobile phases and other LC settings were similar to a screening method for pesticides [67], except for mobile phase B, where methanol was replaced by acetonitrile. To elute all compounds of interest methanol/water/amm.form/FA (95:5 v/v, 2 mM, 0.5 mM) needed to be replaced by acetonitrile/water/amm.form/FA (90:10 v/v, 2 mM, 0.5 mM). The following gradient was tested: 0.1 minute at 10% mobile phase B, then linearly increased to 100% mobile phase B in 12.9 minutes and kept at 100% mobile phase B for 4 minutes. Subsequently the gradient went back to 10% mobile phase B in 0.1 minute and kept at 10% mobile phase B for 2.9 minutes to equilibrate the column. A good separation was obtained, except that the 16:0 OA ester was still retained on the column. To elute the esters the gradient was extended and was kept at 100% mobile phase B for 12 minutes. Although most standards were eluted after 7 minutes the percentage of mobile phase B was slowly increased to obtain a good retention for any unknown phycotoxins eluting in the first 7 minutes. Chromatograms are shown in Figure 3. The final gradient and other LC settings are described in the materials and methods chapter.
H
YDROPHILIC INTERACTION LIQUID CHROMATOGRAPHYFor HILIC methods in general it is known that matrix effect have a major influence on sensitivity and retention times, especially when samples are treated with a non-selective sample treatment [68]. To obtain a sufficient separation of hydrophilic phycotoxins two different HILIC columns were tested. The Nucleoshell HILIC 3 µm, 100 · 2.7 mm HPLC column (Macherey-Nagel, Düren, Germany) and the TOSOH Bioscience TSKgel Amide-80 2 µm, 150 · 3 mm HPLC column (Tosoh Bioscience, Tokyo, Japan) were tested and compared. The best results were obtained with the TSKgel Amide column, as the Nucleoshell HILIC column, which is a zwitterionic column, gave no peaks at all under the tested conditions. However, peaks were still relatively broad with the TSK Amide column and there was no baseline separation for GTX1 and GTX4, which are isomers. Peak shapes of DBA and BMAA were poor and therefore these compounds were excluded from the validation. Chromatograms are shown in Figure 4. The gradient and other LC settings are described in the materials and methods chapter.
28
FIGURE 3: REVERSED PHASE LIQUID CHROMATOGRAPHY OF A) AZASPIRACIDS, SPIROLIDES, PINNATOXINS, GYMNODIMINE AND OKADAIC ACID ESTERS MEASURED IN POSITIVE IONISATION MODE; B) DOMOIC ACID, MICROCYSTINS, CIGAUTOXINS, BREVETOXINS, PECTENOTOXIN AND PALYTOXIN MEASURED IN POSITIVE IONISATION MODE; C) YESSOTOXINS, OKADAIC ACID AND DINOPHYSIS TOXINS MEASURED IN NEGATIVE IONISATION MODE
0% 20% 40% 60% 80% 100% 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 R e lat iv e abundance tR (min) Gym PnTX E 13,19didesmSPXC SPX1 20meSPXG PnTXF PnTXG AZA4 AZA5 OAme-ester OAdiolester AZA3 AZA1 AZA2
A
0% 20% 40% 60% 80% 100% 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 R e lat iv e abundance tR (min) DA NOD MCRR MCHtyR, MCLR, AspMCLR, MCLY MCHilR MCWR, PlTX MCLA MCLY MCLW MCLF pCTX1 PTX2 PbTx3 PbTx9 pCTX2 PbTx2 pCTX3B
0% 20% 40% 60% 80% 100% 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 R e lat iv e abundance tR (min) OA, YTX, hYTX DTX2 DTX1 16:0 OA esterC
29
FIGURE 4: HYDROPHILIC INTERACTION LIQUID CHROMATOGRAPHY OF A) ANATOXIN, CYLINDROSPERMOPSIN, DOMOIC ACID, TETRODOTOXIN, SAXITOXIN AND NEOSAXITOXIN MEASURED IN POSITIVE IONISATION MODE; B) C1&2, GONYAUTOXIN 2, 3 AND 5 MEASURED IN NOEGATIVE IONISATION MODE; C) DECARBAMOYLGONYAUTOXIN 2 AND 3 MEASURED IN NEGATIVE IONISATION MODE; D) GONYAUTOXIN 1 AND 4 MEASURED IN NEGATIVE IONISATION MODE
0% 20% 40% 60% 80% 100% 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 R e lat iv e abundance tR (min) ATX CYN DA TTX STX, dc STX NEO, dcNEO
