Environ. Sci. Technol. 2003, 37, 3078-3084
Geographical, Temporal, and Species Variation of the Polyether Toxins, Azaspiracids, in Shellfish AMBROSE FUREY, CIAN MORONEY, ANA BRAN ˜ A MAGDALENA, M A R I A J O S EÄ F I D A L G O S A E Z , MARY LEHANE, AND KEVIN J. JAMES* PROTEOBIO, Mass Spectrometry Center for Proteomics and Biotoxin Research, Department of Chemistry, Cork Institute of Technology, Bishopstown, Cork, Ireland
Azaspiracid Poisoning (AZP) is a new toxic syndrome that has caused human intoxications throughout Europe following the consumption of mussels (Mytilus edulis), harvested in Ireland. Shellfish intoxication is a consequence of toxin-bearing microalgae in the shellfish food chain, and these studies demonstrated a wide geographic distribution of toxic mussels along the entire western coastal region of Ireland. The first identification of azaspiracids in other bivalve mollusks including oysters (Crassostrea gigas), scallops (Pecten maximus), clams (Tapes phillipinarium), and cockles (Cardium edule) is reported. Importantly, oysters were the only shellfish that accumulated azaspiracids at levels that were comparable with mussels. The highest levels of total azaspiracids (µg/g) recorded to-date were mussels (4.2), oysters (2.45), scallops (0.40), cockles (0.20), and clams (0.61). An examination of the temporal variation of azaspiracid contamination of mussels in a major shellfish production area revealed that, although maximum toxin levels were recorded during the late summer period, significant intoxications were observed at periods when marine dinoflagellate populations were low. Although human intoxications have so far only been associated with mussel consumption, the discovery of significant azaspiracid accumulation in other bivalve mollusks could pose a threat to human health.
Introduction Acute human intoxications, following the consumption of shellfish, can occur due to the contamination by natural toxins that originate in marine microalgae (1, 2). The accumulation of marine toxins by filter-feeding shellfish is a well-known global phenomenon. Bivalve mollusks, including mussels, clams, oysters, and scallops, are the most common vectors of shellfish toxins to humans and the syndromes include diarrhetic (DSP), amnesic (ASP), paralytic (PSP), and neurotoxic (NSP) shellfish poisoning (3-7). The lack of environmental contamination from industrial or large urban populations, together with the temperate climate and rich abundance of phytoplankton, render the western coastline of Ireland as ideal for bivalve shellfish production. However, the rapid expansion of shellfish production has been blighted in recent years by a series of human intoxications due to DSP (8, 9) and a new toxic syndrome, azaspiracid * Corresponding author phone: +353 21 4326317; fax: +353 21 4345191; e-mail:
[email protected]. 3078
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 14, 2003
poisoning (AZP) (10). These events prompted prolonged closures of production areas, together with prohibitions against the importation of Irish shellfish by some European countries. AZP was first identified following two incidents of human intoxications that occurred following the consumption of mussels from Killary and Arranmore Island, Ireland (Figure 1) (11, 10). The human symptoms that included nausea, vomiting, severe diarrhea, and stomach cramps were similar to DSP, but toxins in this class were not detected in significant amounts (12). Azaspiracid (AZA1) induced necroses in the lamina propria of the small intestine, thymus, and spleen and fatty changes in the liver of mice, and these morphopathological changes were distinctly different from other shellfish toxins (13). Chronic toxicological studies using mice showed that azaspiracid produced widespread organ damage and tumors (14). AZA1 is usually the predominant toxin of the AZP class and its unique structural features, characterized by a trispiro ring assembly and an azaspiro ring fused with a 2,9dioxabicyclo[3.3.1]nonane ring, have prompted several groups to explore its synthesis (15, 17). Two analogues, 8-methylazaspiracid (AZA2) and 22-demethylazaspiracid (AZA3), are also found in mussels, usually at lower abundance than AZA1 (10), and AZA4 and AZA5 (18), the hydroxylated analogues of AZA3, are minor contaminants (Figure 2). Several analytical methods involving liquid chromatography, coupled with mass spectrometry (LC-MS) (19-21) have been developed for the determination of azaspiracids. These methods have been applied to the determination of azaspiracids in mussels and phytoplankton (22). LC-MS3 methods, using electrospray ionization with ion-trap mass spectrometry, have recently been used for the determination of the five known azaspiracids in shellfish (23). These methods facilitated the important discovery that these toxins originate in a Protoperidinium sp. This ubiquitous phytoplankton genus was previously regarded as toxicologically benign (24). The AZP syndrome is probably widespread but generally unreported because its symptoms are similar both to DSP and to bacterial enterotoxin poisoning. Recently, azaspiracids were identified in shellfish from England and Norway, indicating a probable widespread risk of AZP in Europe (25). The European Union (EU) has recently imposed a limit of 0.16 µg/g of total AZP toxins in shellfish tissue. The primary aims of this research were to study the geographic and temporal distribution of azaspiracids in mussels in Irish coastal waters and to explore the potential for other types of bivalve mollusks to accumulate azaspiracids.
Experimental Methods Azaspiracid standard toxins, AZA1, AZA2, and AZA3, were isolated from mussels (Mytilus edulis) that were collected from Arranmore Island, Ireland, as described previously (10, 11). Azaspiracid solutions were calibrated using authentic standard AZA1 (10) that was kindly supplied by Prof. M. Satake, Tohoku University, Sendai, Japan. HPLC-grade acetonitrile and water were purchased from Labscan (Dublin, Ireland) and trifluoroacetic acid (TFA) from Sigma-Aldrich (Dorset, U.K.). Contaminated mussels (M. edulis), oysters (Crassostrea gigas), scallops (Pecten maximus), cockles (Cardium edule), and clams (Tapes phillipinarium) were harvested from shellfish production sites throughout Ireland during 1999-2001. Extraction of Azaspiracids from Shellfish. The extraction protocol was based on the procedure that was previously developed, and the efficiency and reproducibility was 10.1021/es020246z CCC: $25.00
2003 American Chemical Society Published on Web 06/17/2003
Xcalibur software. The LC system was a Waters 2690 Alliance (Waters Corporation, Milford, MA) HPLC system that included an autosampler which maintained sample vials at 4 °C. Isocratic chromatography was performed using acetonitrile-water (70:30) containing 0.05% trifluoroacetic acid (TFA) and 0.5 mM ammonium acetate, at a flow rate of 200 µL/min, with a reversed phase precolumn and an analytical column (Luna-2, 5 µm, 30 × 2.0 mm and 150 × 2.0 mm, Phenomenex, Macclesfield, UK) at 40 °C. Using an automated sequence, the eluent flow was diverted to waste for 2 min after sample injection and MS detection was carried out between 2 and 10 min of the chromatography, followed by a second divert to waste for 2 min prior to the next chromatographic sequence.
FIGURE 1. Structures of the three predominant azaspiracids. AZA1 (R1 ) H, R2 ) CH3); AZA2 (R1 ) CH3, R2 ) CH3); AZA3 (R1 ) H, R2 ) H).
Mass spectrometric analysis was carried out in atmospheric pressure ionization using an electrospray source in positive mode. The MS was tuned using azaspiracid standard solution (1 µg/mL) which was infused at 0.3 mL/min with monitoring of the [M + H]+ ion at m/z ) 842. The optimized instrument settings were as follows: capillary temperature (195 °C), spray voltage (5.5 kV), capillary voltage (30 V) and with the arbitrary units for sheath gas flow and auxiliary gas flow set at 90 and 20, respectively. Multiple tandem MS produced collision-induced dissociation (CID) spectra and were obtained using the protonated molecule for each toxin which fragmented similarly giving major ions due to the sequential loss of water molecules. The optimized relative collision energies were 25% for MS2 and 33% for MS3 experiments. Azaspiracids were determined using LC-MS3 with the following target parent and fragment ion combinations in the mass spectrometer: AZA1 (m/z ) 842; 824; 806); AZA2 (m/z ) 856; 838; 820); AZA3 (m/z ) 828; 810; 792).22 Linear calibrations were obtained for AZA1 in shellfish extracts (0.05-1.0 µg/mL) with good regression data, average r2 ) 0.998. Quantitation of AZA2 and AZA3 was based on AZA1 standard calibrations. The limit of detection (S/N ) 3) of AZA1 was better than 10 pg, which is equivalent to 0.9 ng/g shellfish tissue when using the extraction protocol outlined. Good reproducibility data were achieved for AZA1 in shellfish extracts with %RSD values (N ) 5) better than 4.2% (0.05 µg AZA1/mL).
FIGURE 2. cultivation Dunmanus Island; #5: Island; #9:
Map of Ireland showing the locations of shellfish sites where azaspiracids were detected. Site #1: Bay; #2: Bantry Bay; #3: Kenmare Bay; #4: Valentia Killary; #6: Clew Bay; #7: Bruckless; #8: Arranmore Lough Foyle.
established using mussel tissues that were spiked with azaspiracids (22). Tissues (20-100 g) were removed from shellfish (8-18) and were homogenized (Ultra Turrax, IKA, Staufen, Germany) for 1 min. A portion (ca. 5 g) was accurately weighed into a centrifuge tube (50 mL), and acetone (8 mL) was added and again homogenized for 1 min followed by centrifugation at 3000 g for 3 min. The supernatant was transferred to a volumetric flask (25 mL), the extraction procedure was repeated on the residue, and the supernatant was transferred to the volumetric flask that was made up to 25 mL with acetone. Acetone (2.5 mL) was evaporated (TurboVap, Zymark, Warrington, UK), water (0.5 mL) and ethyl acetate (2 mL) were added to the remaining aqueous residue, and vortex was mixed for 1 min. After centrifugation (3000 g), the ethyl acetate layer was removed, and the extraction was repeated. The combined ethyl acetate extracts were evaporated, and the residue was reconstituted in acetonitrile (200 µL). An aliquot (5 µL) was used for LC-MS3 analysis. Liquid Chromatography-Mass Spectrometry (LC-MS). The LC system was linked to a LCQ ion-trap mass spectrometer (ThermoFinnigan, San Jose, CA), controlled using
Results and Discussion Analytical protocols utilizing electrospray LC with multiple tandem MS have become the methods of choice for the determination of trace contaminants present in complex environmental samples. The high sensitivity and selectivity reduces the need for rigorous sample pretreatment and cleanup. A feature of the quadrupole ion-trap MS instrument is the capability of achieving high sensitivity of detection in multiple MS modes (26, 27). In previous studies, it was demonstrated that the determination of toxins in shellfish can be problematic using full-scan MS due to matrix interferences. In full-scan MS mode, signals due to the presence of nontarget components can be observed in chromatograms (22, 28). In multiple tandem MS modes, using an ion-trap instrument, target ions are selected for trapping at each MS stage. The process involves trapping the target ions, clearing the ion-trap of unwanted ions, followed by collision induced fragmentation (MS2) with helium atoms at an optimized energy. The sequence is repeated (MS3) for the product ions which leads to a high selectivity. MS3 has been shown to be more sensitive than single stage MS in an iontrap instrument (22). This is a consequence of the reduction in background noise in multiple tandem MS being more efficient than the decline in analyte signal which results in an improved S/N (26, 27). Not only does LC-MS3 provide excellent quantitative data but confirmation of toxin identity is also accomplished by generating spectra rich in characVOL. 37, NO. 14, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3079
TABLE 1. Summary of Azaspiracid Contamination of Multiple Bivalve Shellfish Species Surveyed shellfish species
sample no.
% positive (>0.01 µg/g)
% exceeding regulatory limita
location no.
maximum AZAs (µg/g)
mussels (M. edulis) oysters (C. gigas) scallop (P. maximus) cockles (C. edule) clams (T. phillipinarium)
800 150 40 20 40
92 67 75 15 15
48 14 50 5 5
S1,S2,S3,S4 S5,S6,S7,S8,S9 S1,S2,S3,S4 S5,S6,S7 S4,S5,S6,S9 S2,S4 S2,S4
4.20 2.45 0.40 0.20 0.61
a
EU regulatory limit ) 0.16 µg (total AZA)/g.
FIGURE 3. Chromatograms from the LC-MS3 analysis of azaspiracids in an extract from mussels (M. edulis). The targeted ions were AZA1 (m/z ) 842; 824; 806); AZA2 (m/z ) 856; 838; 820); AZA3 (m/z ) 828; 810; 792). Elution times were AZA3 (3.29 min), AZA1 (4.23 min), and AZA2 (5.01 min). Chromatographic conditions: Luna-2 C18 column (5 µm, 150 × 2.0 mm) at 40 °C; mobile phase was acetonitrilewater (70:30) containing 0.05% trifluoroacetic acid (TFA) and 0.5 mM ammonium acetate at a flow rate of 200 µL/min. teristic fragment ions. Reproducible retention times were obtained for analytes during chromatographic run sequences of 40-60 analyses. Contamination of the MS was minimized by diverting the chromatographic eluent to waste at the beginning and at the end of each analysis. Geographic Variation of Azaspiracids in Mussels. The reported incidents of human intoxications due to AZP occurred following the consumption of mussels (M. edulis) from small shellfish producing regions of Ireland. Therefore, a study was undertaken to establish the extent of azaspiracid contamination of mussels throughout Ireland, including the major production areas. Mussels were collected over a 3-year period, with an emphasis on sampling during May-October each year, when typically there is the highest risk of marine biotoxin contamination. It was apparent from this study that azaspiracid contamination of mussels occurred along the entire western coastline of Ireland and occasionally at toxicologically dangerous levels. Table 1 shows that AZAs were found in mussels at each of the sampling locations (Figure 2). AZAs were detected (>0.01 µg/g) in 92% of 800 mussel samples examined during this study. The toxin profiles showed that AZA1 was usually the predominant toxin. The toxin profiles in mussels are usually different from other bivalve mollusks, and this will be discussed in detail later. Most positive mussel samples contained both AZA2 and AZA3 that were usually 3080
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 14, 2003
FIGURE 4. Spectra, corresponding to the chromatograms in Figure 3, that were obtained using electrospray ion-trap mass spectrometry by targeting the parent, [M + H]+, and the water-loss ions, [M + H - H2O]+ and [M + H - 2H2O]+. present at lower concentrations than AZA1. For the determination of azaspiracids in shellfish, an LC-MS3 method22 was implemented utilizing the parent and product ions, [M + H]+ f [M + H - H2O]+ f [M + H - 2H2O]+. Nine ions, representing the three major azaspiracids, AZA1-AZA3, were trapped using narrow mass windows, rendering the method very selective. The limit of detection (L.O.D.) was better than 10 pg for AZA1-AZA3, which corresponds to 0.9 ng/g tissue. Figure 3 shows typical chromatograms obtained for AZA1-AZA3 using a reversedphase column under isocratic conditions. The corresponding spectra are shown in Figure 4 showing characteristic
FIGURE 5. Representative chromatograms obtained using LC-MS3 analysis of extracts from four different shellfish species. Scallop (P. maximus): AZA1 (82%), AZA2 (18%); oyster (C. gigas): AZA1 (72%), AZA2 (28%); clam (T. phillipinarium): AZA1 (75%), AZA2 (25%); cockle (C. edule): AZA1 (44%), AZA2 (56%). See Figure 3 for chromatographic conditions. backbone fragmentation of azaspiracids (22) commencing with fracture of the A-ring: AZA1 (m/z ) 672); AZA2 (m/z ) 672); AZA3 (m/z ) 658). The azaspiracid toxin profile in mussels was similar to those found in trawl samples containing mixed phytoplankton and in monocultures of Protoperidinium spp (24). The highest concentration of AZAs found in mussels was 4.2 µg total AZAs/g, comprising AZA1 (2.2 µg/g), AZA2 (1.4 µg/g), AZA3 (0.60 µg/g). Also, the fact that during the period of this study, a high proportion of mussels (48%) contained AZAs in excess of the regulatory limit poses problems for sustainable commercial mussel cultivation. The Identification of Azaspiracids in Multiple Shellfish Species. Since marine biotoxins are not usually confined to a single shellfish species, this study explored the extent to which other bivalve mollusks could accumulate azaspiracids. In addition to mussels, four other species of bivalve mollusks,
oysters (C. gigas), scallops (P. maximus), cockles (C. edule), and clams (T. phillipinarium) were examined and found to be positive for AZAs. All of the shellfish species are commercially cultivated and exported to many countries in Europe. The discovery that other bivalve shellfish are potential azaspiracid toxin carriers is therefore important for food safety. Although human intoxications due to AZP have so far been reported only following the consumption of mussels, the significant toxicity that was identified in other shellfish species confirms that they also need to be monitored. Azaspiracids in Oysters. Studies of shellfish intoxications due to DSP toxins seldom showed significant contamination of oysters even when mussels from the same region contained high toxin levels (unpublished data). This is attributed mainly to different cultivation methods as oysters are cultivated along the shoreline and mussels are grown on suspended ropes in deeper waters. One of the main conclusions of the current VOL. 37, NO. 14, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3081
FIGURE 6. Temporal variation of azaspiracids in mussels (M. edulis) from a cultivation site (#2, Figure 2) commencing September 1999 and ending August 2000. (A) Azaspiracid toxin profile in the digestive glands of mussels. (B) Azaspiracid toxin profile in the remaining mussel tissues. study was that both oysters and rope-cultivated mussels are similarly susceptible to contamination by AZAs. In some instances, the levels found in oysters were up to 5-fold higher. In sample site #7, County Donegal, where both mussels and oysters were cultivated, 2.45 µg total AZA/g were found in oysters which is 15 times the regulatory limit. The toxin profile was AZA1 (1.55 µg/g), AZA2 (0.8 µg/g), AZA3 (0.1 µg/g) and 3082
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 14, 2003
mussels that were examined from the same site on the same date contained 1.8 µg total AZAs/g, comprising AZA1 (1.1 µg/g), AZA2 (0.23 µg/g), AZA3 (0.45 µg/g). Figure 5 shows representative chromatograms that were obtained for four shellfish species. An interesting observation was that AZA3 was usually absent from oysters or present at much lower levels than in mussels. Although most of the 150 oyster
samples examined contained detectable levels of AZAs, only 14% exceeded the regulatory limit. We conclude from this study that oysters can be efficient accumulators of AZAs and, although the human health risk is less than mussels, the contamination is sufficiently serious to require vigilant monitoring of oysters for azaspiracids. Azaspiracids in Clams, Cockles, and Scallops. There was much lower frequency of detection of AZAs in clams (T. phillipinarium) and cockles (C. edule) (Table 1) with only 5% of samples exceeding the regulatory limit. However, one sample of clams contained 0.61 µg AZAs/g which was significantly in excess of the safe limit. The toxin profile was AZA1 (0.46 µg/g) and AZA2 (0.15 µg/g) but no detectable level of AZA3. The highest concentration of AZP toxins found in cockles was 0.20 µg AZAs/g, comprising AZA1 (0.08 µg/g), AZA2 (0.12 µg/g), AZA3 (0.00 µg/g). The level of contamination in this sample therefore marginally exceeded the limit. This study has demonstrated that cockles and clams are capable of accumulating azaspiracids. Although it is probable that these species pose a much lower risk for human consumers than mussels and oysters, their inclusion in monitoring programs is prudent. The highest concentration of toxins found in scallops (P. maximus) was 0.40 µg AZAs/g, comprising only AZA1 and AZA2 (0.20 µg/g), but AZA3 was not detected (Figure 5). Most cultures consume only part of scallop tissues, usually the adductor muscle (meat) and gonad. In this regard, toxicological assessments relating to scallops are somewhat different from the other bivalve shellfish that are consumed whole. Examination of the various tissue compartments of scallops revealed that contamination of the adductor muscle and gonad by AZAs was consistently low,
