ARTICLE pubs.acs.org/est
Mussels Increase Xenobiotic (Azaspiracid) Toxicity Using a Unique Bioconversion Mechanism Daniel O’Driscoll,† Zuzana Skrabakova,†,‡ John O’Halloran,‡,§ Frank N. A. M. van Pelt,‡,|| and Kevin J. James*,‡ †
PROTEOBIO (Mass Spectrometry Centre), Cork Institute of Technology, Bishopstown, Cork, Ireland Environmental Research Institute, University College Cork, Lee Road, Cork, Ireland § School of Biological, Earth and Environmental Sciences, Enterprise Centre, Distillery Fields, University College Cork, Ireland Department of Pharmacology and Therapeutics, University College Cork, Cork, Ireland
)
‡
bS Supporting Information ABSTRACT: Azaspiracid Poisoning (AZP) is a human toxic syndrome which is associated with the consumption of bivalve shellfish. Unlike other shellfish, mussels contain a large array of azaspiracid analogs, many of which are suspected bioconversion products. These studies were conducted to elucidate the metabolic pathways of azaspiracid (AZA1) in the blue mussel (Mytilus edulis) and revealed that the main biotransformation product was the more toxic demethyl analog, AZA3. To elucidate the mechanism of this C-demethylation, an unprecedented xenobiotic bioconversion step in shellfish, AZA1 was fed to mussels that contained no detectable azaspiracids. Triple quadrupole mass spectrometry (MS) and high resolution Orbitrap MS were used to determine the uptake of AZA1 and the toxin profiles in three tissue compartments of mussels. The second most abundant bioconversion product was identified as AZA17, a carboxyl analog of AZA3, which is a key intermediate in the formation of AZA3. Also, two pairs of isomeric hydroxyl analogs, AZA4/AZA5 and AZA7/AZA8, have been confirmed as bioconversion products for the first time. Ultra high resolution (100 k) MS studies showed that the most probable structural assignment for AZA17 is 22-carboxy-AZA3 and a mechanism for its facile decarboxylation to form AZA3 has been proposed.
’ INTRODUCTION Azaspiracids (AZAs) are polyether marine toxins, produced naturally by marine phytoplankton. These toxins bioconcentrate in bivalve mollusks, such as mussels, clams, oysters, and scallops, that filter-feed on phytoplankton. The first confirmed human intoxication incident involving AZAs was in The Netherlands in 1995, following reports of severe gastrointestinal illness after consuming mussels (Mytilus edulis) that were cultivated on the west coast of Ireland. Azaspiracid (AZA1) was isolated from these shellfish, and the new toxic syndrome was named Azaspiracid Poisoning (AZP).1,2 The initially proposed structure for AZA1 was subsequently modified following a total synthesis.3 Other food-borne illnesses involving AZAs were also associated with the consumption of mussels cultivated in Ireland.4 Although AZA1 is the most abundant toxin from this group, AZA2 and AZA3, the 8-methyl and 22-demethyl analogs of AZA1, respectively, frequently co-occur in mussels.2,5,6 To date, more than 20 analogs have been detected in shellfish but only AZA1-AZA5 (Figure 1, Table 1) have been fully structurally elucidated.7 r 2011 American Chemical Society
The rapid development of highly sensitive liquid chromatography - multiple tandem mass spectrometry (LC-MS/MS, LCMSn) methods were invaluable to the control of shellfish contamination.8,9 AZAs have been identified in shellfish from several other European countries.10-12 These toxins have also recently been reported in the coastal waters of Chile13 and Japan.14 Azaspiracids have so far been identified in two dinoflagellate species, Protoperidinum crassipes15and Azadinium spinosum,16 and in a sponge (Echinochlathria sp.).14 AZA1 is a potent cytotoxin, and studies using mice have shown that this toxin causes widespread organ damage.17 AZA2 and AZA3 are more toxic than AZA1, and the target organs include the liver, spleen, and small intestine.17,18 A chronic study using sublethal doses of AZA1 administered to mice showed tumor formation in the lungs as well as malignant lymphomas.19 Received: October 26, 2010 Accepted: February 24, 2011 Revised: February 22, 2011 Published: March 14, 2011 3102
dx.doi.org/10.1021/es103612c | Environ. Sci. Technol. 2011, 45, 3102–3108
Environmental Science & Technology
ARTICLE
Table 1. Designation of Azaspiracid Substituents, R1, R2, R3, and R4 toxin
R1
R2
R3
R4
[MþH]þ m/z
AZA1
H
CH3
H
H
842.50545
AZA2 AZA3
CH3 H
CH3 H
H H
H H
856.52110 828.48980
AZA4
H
H
OH
H
844.48472
AZA5
H
H
H
OH
844.48472
AZA7
H
CH3
OH
H
858.50037
AZA8
H
CH3
H
OH
858.50037
AZA17
H
COOH
H
H
872.47963
Although other bivalve shellfish (oysters, scallops, cockles, and clams) and crustaceans can accumulate azaspiracids, only mussels have so far been associated with AZP in humans.4 We previously reported that only AZA1 and AZA2 are found in most shellfish species, but mussels differ in containing a much more complex toxin profile with multiple AZA analogs frequently present.6 The main objective of this investigation was to identify the main AZA1 bioconversion products and pathways in mussels (M. edulis). The toxin composition in the tissue compartments of mussels was determined using LC-MS/MS and a recently developed ultra high resolution (100 k fwhm) full-scan MS (Orbitrap) method to achieve non target analysis of biotransformation products of AZA1.20
’ EXPERIMENTAL SECTION Toxin Standards. Toxic mussels (Mytilus edulis) were collected from the northwest coast of Ireland and azaspiracids (AZA1-AZA8 and AZA17) were isolated using procedures that were previously developed.2,21 Calibration studies were performed using a certified reference standard AZA1, CRM-AZA1 (1.24 μg/mL; NRC, Halifax, Canada). The standard solutions of AZA1 and all of the other azaspiracids were quantified with this certified reference AZA1 using high resolution full-scan MS. The standard solutions of azaspiracids were used for the construction of calibration curves. Mussel Acclimation and AZA1 Feeding Experiments. Mussels (M. edulis) were collected from County Waterford, Ireland, and acclimated to laboratory conditions for 12 months. During this period, mussels were kept in two 100 L tanks containing 60 L of seawater, maintained at 14 ( 0.5 °C, with constant aeration, under low light intensity exposure and fed with a shellfish diet (described below). One fifth of the seawater volume was replaced every day. Mussels and seawater were analyzed using LC-MS/MS to ensure that they were free of azaspiracids and other marine toxins, including dinophysistoxins and pectenotoxins. For feeding experiments, 10 mussels were placed in each of two experimental glass tanks (20 L), containing seawater (10 L) at 14 °C. Shellfish diet 1800 (a mixed diet of Isochrysis, Pavlova Thalassiosira weissfloii Tetraselmis) was administered in accordance with the supplier’s recommendations (Reed Mariculture Inc., Cambell, CA, USA). An aliquot (50 mL) of seawater from the tank was removed for toxin analysis every hour over a 12-h period for determination of the rate of toxin uptake and at 24-h before implementation of the 100% seawater change. Pure AZA1 (3 μg/day), mixed with the shellfish diet, was administered over a 10 day period. After day 10, six mussels were removed and
Figure 1. General structure of azaspiracids. (See Table 1 for the designations of R1, R2, R3, and R4.)
opened, rinsed in Milli-Q water, and weighed. The shellfish were dissected into three parts (digestive glands or hepatopancreas (HP), gills, and remaining tissues), and an extraction procedure was carried out on each tissue compartment. A complete seawater change was carried out with feeding but without toxin administration, and this was repeated each day for the remaining three days. The remaining mussels were removed after thirteen days, and their tissues were analyzed for azaspiracids. Aliquots (50 mL) were also taken from the experimental tank for toxin analysis. Extraction of Azaspiracids from Mussel Tissues. Mussel tissue (M. edulis) was divided into HP, gills, and remaining tissue and were analyzed separately. Shellfish tissue was combined with acetone (4 mL) in a 50 mL centrifuge tube and homogenized (Ultra Turrax, IKA, Germany) for 1 min followed by centrifugation at 3000 rpm (700 g) for 3 min. The supernatant was transferred to a centrifuge tube (50 mL), the extraction procedure was repeated on the residue, and the supernatant again was combined with the above to give a final volume of 8 mL. An aliquot (3 mL) was evaporated using nitrogen at 40 °C to remove acetone (TurboVap, Zymark, Hopkinton, MA, USA). To the remaining aqueous residue, water (0.5 mL) and ethyl acetate (2 mL) were added with vortex mixing for 1 min. The solution was centrifuged at 700 g for 3-4 min, and the ethyl acetate was transferred to a glass tube (10 mL). The extraction with ethyl acetate (2 mL) was repeated, and the combined extracts were evaporated using nitrogen. The residue was reconstituted with acetonitrile (200 μL) using sonication and vortex mixing, and 35 μL of this solution was used for LC-MS/MS analysis. Liquid Chromatography - Mass Spectrometry (LC-MS/MS). Triple Stage Quadrupole Mass Spectrometry. An API 3000 (Applied Biosystems, Warrington, UK) mass spectrometer (triple stage quadrupole) with a Turbo Ionspray interface operating in positive mode was coupled to an Agilent 1100 series LC system (Agilent, Palo Alto, CA, USA). The software used was Analyst 1.5 for data acquisition, data analysis, and instrument control. The ion source dependent parameters were optimized as follows: nebulizer gas 10 (arb), curtain gas 15 (arb), Ionspray voltage (IS) 4250 V, temperature (TEM) 550 °C, and collision gas (CAD) was set to 5 (arb). Multiple reaction monitoring (MRM) experiments were developed using voltages optimized for pseudomolecular ion production: declustering potential (DP) 60 V, focusing potential (FP) 300 V, entrance potential (EP) 10 V, collision cell exit potential (CXP) 15 V. Two collision 3103
dx.doi.org/10.1021/es103612c |Environ. Sci. Technol. 2011, 45, 3102–3108
Environmental Science & Technology energies were adopted for two transitions of 14 azaspiracids (AZA1AZA12, AZA17, and AZA19). The [MþH]þ f [MþH - H2O]þ transition, collision energy (45 V), was used for quantitation of each azaspiracid except AZA17 and AZA19 for which the [MþH]þ f [MþH - H2O-CO2]þ transitions were used (collision energy 54 V). The [MþH]þ f [MþH - H2O-A-ring]þ transition, collision energy (70 V), was used for confirmation. Chromatographic separation of azaspiracids was achieved by using isocratic elution (acetonitrile/water (58:42) v/v containing 0.05% formic acid) on a reversed phase column (Luna-C-18 (2), 5 μm, 150 2.0 mm; Phenomenex, Macclesfield, UK), at 35 °C. Samples (3 -5 μL) were injected, and the eluent flow was directed to waste for 1 min. The flow rate was 200 μL/min, and the total chromatographic time was 25 min. LTQ Orbitrap Mass Spectrometry. An Accela LC system (Thermo Scientific, Hemel Hempstead, UK) was used. The analytical column used was a Luna C18(2) (3 μm, 150 2 mm), equipped with a SecurityGuard cartridge (C18, 4.0 mm 2.0 mm, Phenomenex, Macclesfield, UK), at 35 °C. Isocratic elution was used; water (42% v/v) and acetonitrile (58% v/v), both containing 0.05% trifluoroacetic acid (TFA) and ammonium acetate (1 mM), at a flow rate of 200 μL/min. After sample (5 μL) injection, the eluate from the column was diverted to waste for one min and the chromatography was terminated at 15 min. The LC system was connected to a hybrid linear ion-trap (LIT) Fourier transform (FT) mass spectrometer (LTQ Orbitrap XL, Thermo Scientific), fitted with a heated electrospray interface (H-ESI) and operating in positive ionization mode. All ion source tune parameters were optimized manually by infusing AZA1 standard (0.5 μg/mL) and monitoring the [MþH]þ ion at m/z 842.5. This optimization was carried out by adjusting various parameters until no precursor ion fragmentation was observed. For the determination of azaspiracids, the method consisted of two scan events: a) full-scan FTMS (Orbitrap XL) and b) FTMS HCD MS/MS data-dependent scans on selected ions (Table 1). The mass resolution was set at 100,000 (fwhm) with a mass range of m/z 100-900 in the Orbitrap mass analyzer. The precursor ion chromatograms were obtained from full-scan data using a mass tolerance window of (2 mDa. The lock mass option was employed by recalibration of the mass scale using the m/z value of the [MþNa]þ ion from TFA. The full validated analytical procedure for the determination of azaspiracids in mussel tissues was carried out as described in Skrabakova et al.20
’ RESULTS AND DISCUSSION AZA1 Feeding Experiments. The feeding experiments were designed to efficiently administer AZA1 to mussels (M. edulis) so that the distribution of this toxin and its bioconversion products, in three mussel tissue compartments, could be determined using sensitive LC-MS/MS and ultra high resolution full-scan LC-MS methods. Mussels were maintained in a tank (10 L) of seawater at 14 °C, and the study was conducted over a 13-day period. A preliminary experiment was performed to determine the rate of AZA1 uptake by mussels. After administering an algal feed containing AZA1, an aliquot (50 mL) of tank water was removed for toxin analysis. This showed that most of the AZA1 (>95%) had been taken up by mussels in 5 h. AZA1 (30 μg) was administered to M. edulis (10 shellfish) for the bioconversion study at the rate of 3 μg/day, by mixing the toxin with an algal feed. One day after each toxin administration, the water was changed, and an aliquot (50 mL) was retained for toxin analysis.
ARTICLE
Figure 2. Sample chromatograms showing the azaspiracid composition in mussel (M. edulis): a) gills, b) digestive glands, and c) remaining tissues, after feeding AZA1.
After 10 days, six shellfish were removed, dissected into three parts, digestive glands, gills, and remaining tissues (predominantly posterior adductor muscle and mantle). The 18 tissue samples were individually homogenized and extracted immediately. Targeted azaspiracids (14 analogs) were determined in each sample by LC-MS/MS using a triple quadrupole mass spectrometer. The multiple reaction monitoring (MRM) method employed monitored two precursor-product ion transitions with optimized collision energies for 14 azaspiracids (AZA1AZA12, AZA17, and AZA19). The remaining four mussels were administered algal food daily, together with a total seawater change, for another three days after which they were similarly dissected and analyzed for azaspiracids. Examples of extracted ion chromatograms that were obtained are shown in Figure 2. Full quantitation of azaspiracids in shellfish tissues was performed using ultra high resolution full-scan Orbitrap MS.20 AZA1 Bioconversion Products in Mussels. Figure 3 shows the main AZA toxin profiles in three mussel tissue compartments a) after 10 days with daily AZA1 feeding, and b) after 13 days (three additional days with no AZA1 feeding). AZA3 was the predominant toxin in each tissue compartment, and significant levels of AZA17 (22-carboxy-AZA3) were also present in all samples. Full toxin profile data are presented in Supplementary 3104
dx.doi.org/10.1021/es103612c |Environ. Sci. Technol. 2011, 45, 3102–3108
Environmental Science & Technology
ARTICLE
Figure 3. Charts showing the distribution of AZA1, AZA3, and AZA17 after feeding AZA1 to mussels (M. edulis) a) after 10 days and b) after 13 days (no further feeding of AZA1).
Scheme 1. Proposed Mechanism for the Biotransformation of AZA1 to AZA3 That Involves Oxidation and Decarboxylation
Tables 1 and 2. The main differences between the two data sets were a decrease in the combined AZA1, AZA3, and AZA17 content in the gills and an increase of AZA3 content in the remaining mussel tissues. These differences may be due to toxin migration or further biotransformations. The most significant conclusion from these studies is that this is the first proof that AZA1 is converted to AZA3 in mussels, a remarkable C-demethylation. Further, the predominant intermediate in this biotransformation is a carboxyl analog that has been identified as AZA17 (see Scheme 1). Other azaspiracids that were detected included the 3- and 23-hydroxy analogs of AZA1 (AZA7, AZA8, respectively) and of AZA3 (AZA4, AZA5, respectively). AZA4/AZA5 were found at concentration levels less than 2% of total azaspiracids, and even lower concentration levels (
