Tetrodotoxins Occurrence in Non-Traditional Vectors of the North Atlantic Waters (Portuguese Maritime Territory, and Morocco Coast)

Tetrodotoxin (TTX) is a potent alkaloid typically from tropical ecosystems, but in the last decade its presence has been more pronounced in the temperate waters of the Atlantic. In its last scientific opinion, the European Food Safety Authority (EFSA) stressed the need for data regarding TTX prevalence in European waters. To address EFSA’s concerns, benthic organisms such as mollusks, crustaceans, echinoderms and fish with different feeding habits were collected along the Portuguese continental coast, islands (São Miguel, Azores, and Madeira) and the northwestern Moroccan coast. A total of 165 samples were analyzed by ultra high performance liquid chromatography high resolution mass spectrometry (UHPLC-HRMS) and ultra high performance chromatography mass spectrometry (UHPLC-MS/MS). Geographical tendencies were detected as follows, by descending order: S. Miguel Island (Azores), Moroccan coast, Madeira Island and Portuguese continental coast. The toxin amounts detected were significant, above the Dutch limit value established in 2017, showing the importance and the need for continuity of these studies to gain more knowledge about the prevalence of these toxins, unraveling new vectors, in order to better assess human health risk. This work represents a general overview of new TTX bearers (7) most of them in gastropods (Patella depressa, Nucella lapillus, Onchidella celtica and Aplysia depilans), followed by echinoderms (Echinus esculentus and Ophidiaster ophidianus) and puffer fish Sphoeroides marmoratus.


Introduction
Tetrodotoxin (TTX) is an extremely potent alkaloid typical from warm ecosystems that over recent years became more frequent in temperate waters [1][2][3]. Named after the Tetraodontidae puffer fish family where it was first isolated, TTX was later identified also in different taxa not close related (from bacteria; marine invertebrates; terrestrial and marine vertebrates) [4,5]. The ubiquity of TTX is due to its exogenous origin, bacteria from the different phyla (Proteobacteria, Firmicutes, Bacterioides and Actinobacteria) associated with dinoflagellate blooms of Prorocentrum have been pointed to as potential producers [6][7][8].
Chemically, TTX is a crystalline weakly basic heterocyclic molecule with six hydroxyl groups, a pyridine ring with additional fused ring systems and a guanidinium group, positively charged at a physiological pH, with a molecular formula of C 11 H 17 O 8 N 3 (Table 1) [1]. This powerful alkaloid exhibits its action by binding selectively and extracellularly to receptor-site 1 of voltage-gated sodium channels (Na v s), once it mimics the sodium hydrated cation, occluding the outer pore. Thus, avoiding the access of monovalent cations to the pore, the generation and propagation of the action potential is blocked, inhibiting the communication between electrically excitable cells in muscle and nerve tissues, causing paralysis which results in death by cardio-respiratory failure [9][10][11]. About 30 analogues have been described to date, whose toxicity varies with structure implying their greater or lesser affinity for their molecular target [12]. The rise of water temperature together with anthropological intervention aided the migration and establishment of this neurotoxin typically from tropical ecosystems, the Pacific and Indian oceans, into European waters, where it has become more prevalent in the last decade [13][14][15][16][17]. Regarding consumer protection in the EU, regulations 853/2004/EC and 854/2004/EC were published in 2004, preventing fish species reported as TTX-bearers (Tetraodontidae, Canthigasteridae, Molidae and Diodontidae) from market placement [18,19]; the Netherlands being the pioneer in including the scrutinizing of this toxin in its monitoring program, establishing 44 µg TTX equivalents/kg of shellfish as a safe concentration [20]. Still, more data are needed in order to better protect consumer health. In EFSA's latest opinion it's stated that more occurrence data on TTX and its derivatives in edible parts of gastropods and bivalves are necessary in order to complement a reliable exposure assessment, since it was not possible to include gastropods in the risk characterization due to insufficient studies [21]. The primary aim of this work was to screen the presence of TTX in different benthic marine species along the Portuguese continental and archipelagos and western Moroccan coasts, thus unraveling its prevalence and geographical trends. We hope to bridge and to contribute to EFSA's recommendations, towards the development of effective and robust monitoring programs. necessary in order to complement a reliable exposure assessment, since it was not possible to include gastropods in the risk characterization due to insufficient studies [21]. The primary aim of this work was to screen the presence of TTX in different benthic marine species along the Portuguese continental and archipelagos and western Moroccan coasts, thus unraveling its prevalence and geographical trends. We hope to bridge and to contribute to EFSA's recommendations, towards the development of effective and robust monitoring programs.

Results and Discussion
A total of 165 samples collected from 25 different sampling points ( Figure 1) were screened using the techniques of ultra high performance chromatography mass spectrometry (UHPLC-MS/MS) and ultra high performance liquid chromatography high resolution mass spectrometry (UHPLC-HRMS) [8,22]. These two techniques are complementary when performing a complete and undoubted identification of TTX analogues when standards are not available. UHPLC-MS/MS in product ion scan (PIS) mode allowed to check the mass fragmentation pattern of this family of compounds and HRMS was used to obtain the exact mass of each compound. Finally, the positive samples were quantified by UHPLC-MS/MS in multiple reaction monitoring (MRM) mode using a TTX standard. About 30 species of edible, with commercial interest, and non-edible benthic organisms were collected in the intertidal zone and by Self-Contained Underwater Breathing Apparatus (SCUBA) diving, with the aim of searching new vectors and determining the prevalence of TTXs in the food web: bivalves (mussels), gastropods (limpets, sea-snails, sea-slugs), echinoderms (starfish, seaurchins, sea-cucumbers), crustaceans (barnacles) and fish (puffer). Sampling points and collected species are described in detail in the Materials and Methods section.

Results and Discussion
A total of 165 samples collected from 25 different sampling points ( Figure 1) were screened using the techniques of ultra high performance chromatography mass spectrometry (UHPLC-MS/MS) and ultra high performance liquid chromatography high resolution mass spectrometry (UHPLC-HRMS) [8,22]. These two techniques are complementary when performing a complete and undoubted identification of TTX analogues when standards are not available. UHPLC-MS/MS in product ion scan (PIS) mode allowed to check the mass fragmentation pattern of this family of compounds and HRMS was used to obtain the exact mass of each compound. Finally, the positive samples were quantified by UHPLC-MS/MS in multiple reaction monitoring (MRM) mode using a TTX standard. About 30 species of edible, with commercial interest, and non-edible benthic organisms were collected in the intertidal zone and by Self-Contained Underwater Breathing Apparatus (SCUBA) diving, with the aim of searching new vectors and determining the prevalence of TTXs in the food web: bivalves (mussels), gastropods (limpets, sea-snails, sea-slugs), echinoderms (starfish, seaurchins, sea-cucumbers), crustaceans (barnacles) and fish (puffer). Sampling points and collected species are described in detail in the Materials and Methods section. Tetrodonic acid ------------C11H17N3O8 320.1088

Results and Discussion
A total of 165 samples collected from 25 different sampling points ( Figure 1) were screened using the techniques of ultra high performance chromatography mass spectrometry (UHPLC-MS/MS) and ultra high performance liquid chromatography high resolution mass spectrometry (UHPLC-HRMS) [8,22]. These two techniques are complementary when performing a complete and undoubted identification of TTX analogues when standards are not available. UHPLC-MS/MS in product ion scan (PIS) mode allowed to check the mass fragmentation pattern of this family of compounds and HRMS was used to obtain the exact mass of each compound. Finally, the positive samples were quantified by UHPLC-MS/MS in multiple reaction monitoring (MRM) mode using a TTX standard. About 30 species of edible, with commercial interest, and non-edible benthic organisms were collected in the intertidal zone and by Self-Contained Underwater Breathing Apparatus (SCUBA) diving, with the aim of searching new vectors and determining the prevalence of TTXs in the food web: bivalves (mussels), gastropods (limpets, sea-snails, sea-slugs), echinoderms (starfish, seaurchins, sea-cucumbers), crustaceans (barnacles) and fish (puffer). Sampling points and collected species are described in detail in the Materials and Methods section.

Results and Discussion
A total of 165 samples collected from 25 different sampling points ( Figure 1) were screened using the techniques of ultra high performance chromatography mass spectrometry (UHPLC-MS/MS) and ultra high performance liquid chromatography high resolution mass spectrometry (UHPLC-HRMS) [8,22]. These two techniques are complementary when performing a complete and undoubted identification of TTX analogues when standards are not available. UHPLC-MS/MS in product ion scan (PIS) mode allowed to check the mass fragmentation pattern of this family of compounds and HRMS was used to obtain the exact mass of each compound. Finally, the positive samples were quantified by UHPLC-MS/MS in multiple reaction monitoring (MRM) mode using a TTX standard. About 30 species of edible, with commercial interest, and non-edible benthic organisms were collected in the intertidal zone and by Self-Contained Underwater Breathing Apparatus (SCUBA) diving, with the aim of searching new vectors and determining the prevalence of TTXs in the food web: bivalves (mussels), gastropods (limpets, sea-snails, sea-slugs), echinoderms (starfish, sea-urchins, sea-cucumbers), crustaceans (barnacles) and fish (puffer). Sampling points and collected species are described in detail in the Materials and Methods section.
In the present study, a variety of samples from different animals were analyzed using the UHPLC-MS/MS method after the extraction procedure. Initially, samples were analyzed in MRM mode for 18 TTX analogues described in the literature [1] and some compounds of this family were detected in several samples. This preliminary analysis showed many positive samples for TTX plus analogues. One of the great challenges of the emerging toxins is the lack of reference material, as far as TTX is concerned, therefore only the analogues present in the standard (TTX, 4,9-anhydroTTX, 4-epiTTX and 11-deoxyTTX) could be confirmed ( Figure 2). However, all positive extracts presented several TTX analogues. To confirm these positive results, their mass fragmentation pattern was studied [8,23,24]. The fragmentation pathway of TTX standard shows two losses of water and m/z 256, m/z 178 and m/z 162 as product ions ( Figure 3) [8,24,25].  In the present study, a variety of samples from different animals were analyzed using the UHPLC-MS/MS method after the extraction procedure. Initially, samples were analyzed in MRM mode for 18 TTX analogues described in the literature [1] and some compounds of this family were detected in several samples. This preliminary analysis showed many positive samples for TTX plus analogues. One of the great challenges of the emerging toxins is the lack of reference material, as far as TTX is concerned, therefore only the analogues present in the standard (TTX, 4,9-anhydroTTX, 4-epiTTX and 11-deoxyTTX) could be confirmed ( Figure 2). However, all positive extracts presented several TTX analogues. To confirm these positive results, their mass fragmentation pattern was studied [8,23,24]. The fragmentation pathway of TTX standard shows two losses of water and m/z 256, m/z 178 and m/z 162 as product ions ( Figure 3) [8,24,25].       In this way, TTXs analogues were confirmed in several samples. As shown in Table 2, a loss of water and m/z 178 (or 176) and m/z 162 are the product ions of TTX analogues detected in positive samples. At the same time, the suspicious positive samples were also analyzed using High Resolution Mass Spectrometry (HRMS) to confirm and identify TTX analogues. The mass fragmentation pathway of TTX analogues has been reported in many LC-HRMS studies [8,26]. In the current study the exact mass of TTX analogues was determined by HRMS using the formula predictor software to characterize and predict the molecular formula of these unknown compounds. In this way, elemental composition, theoretical value m/z, experimental m/z and mass errors in mDa for TTX analogues were obtained after MS 1 spectra (Tables 3-5). Once the presence of TTX analogues was verified, the toxin amount was quantified using LC-MS/MS in MRM mode (Tables 3-5). In order to calculate the concentrations of TTXs, the calibration curve was done with TTX standard, assuming that related analogues would give a similar response of TTX. The quantification was done using MRM acquisition in positive mode. A seven-point calibration curve with range between 1.5 ng/mL-100 ng/mL was used (R 2 = 0.999). The limit of detection (LOD) was 1.64 µg/Kg and the limit of quantification (LOQ) 5.46 µg/Kg.

Portuguese Continental Coast
In a total of 62 analyzed samples collected between September 2011 and January 2012 at the 11 sampling points (Figure 1), 8 positive results were detected, all below the LOQ (Table 3).
Regarding positive hits for TTXs, despite not quantifiable, we count three first reports of this group of toxins, two in gastropods, P. depressa and N. lapillus, for the analogs 11-oxoTTX and trideoxyTTX, respectively and one in echinoderms, E. esculentus positive for 11-oxoTTX. In our former work, the analogue trideoxyTTX was identified only in one sample of the species C. lampas [16]; in the present study, this analogue was the most common along the Portuguese coast, being present in five of seven positive organisms, all gastropods: P. lineatus, C. lampas, N. lapillus, G. umbilicalis.
Finally, the samples were quantified by MRM using TTX standard, assuming that the TTX analogues have a similar response. In this case, the quantification was not possible since the amount present in the samples was below the LOQ.

Madeira Island (Madeira Archipelago)
From a total of 23 samples harvested in two sampling points (Figure 1) during the summer of 2012, we reported zero positive hits for TTX and its derivatives.

São Miguel Island (Azores Archipelago)
São Miguel island, Azores, was screened during June 2013. From the 39 harvested samples along 6 sampling points (Figure 1), we detected 4 positive hits (Table 4). Regarding toxin uptake, we found in São Miguel, Azores, a distinct toxin profile concerning positive taxa and detected toxins, also all detected values were above the current limit established in the Netherlands [20]. Starting with echinoderms, all positive samples were in the red velvet starfish, O. ophidianus, for an unknown TTX analogue since the molecule that follows TTX mass fragmentation pattern, further studies are needed. The sample of S. marmoratus was analyzed per tissue (muscle, gonads and liver), since the specimen belongs to the Tetraodontidae family, and there was sufficient extractable biomass. Discussing uptake by tissue, the values detected curiously increased in the following order: liver, muscle, gonads. Surprisingly, the liver had half the values detected in the muscle and ten times less than in the gonads, which deviates from the typical accumulation pattern for this family of fish [34,35]. However, a study performed in the United States of America, between 2002 and 2004, showed the same uptake pattern, for the saxitoxin group, in the species Sphoeroides nephelus, detecting concentrations in the muscle 5 to 20 fold higher in comparison to the liver [36]. Concerning toxic profile, all tissues were positive for TTX, 4-epiTTX, monodeoxyTTX, 4,9-anhydroTTX, dideoxyTTX, trideoxyTTX and anhydrotrideoxyTTX, being the later derivative not detected in the liver tissue.
In comparison with other works regarding TTX concentrations in starfish [37,38], though they are a distinct ecotype and species, the present measured concentrations are higher. Though chronologically distant from the Japanese studies, in the present work, the considerable detected amounts could be explained by the rise of water temperature. In a simplistic approach, it has been shown that seawater temperature warming correlates with the global warming phenomenon; it was also proved that TTX production has a positive relationship with water temperature [39]. Other studies concerning other emergent toxin, ciguatera, predict that global warming is expected to be the catalyst for the increment of the incidence and prevalence of poisoning cases [40,41]. From our first report in 2012 on this neurotoxin group in Portugal, Angeiras was the northernmost point of TTX in Europe [16]. The latest reports corroborate this trend, since the last work points to its presence in England [42]. It is noteworthy, that this is the first report of TTXs in both species: O. ophidianus and S. marmoratus.
Three samples of O. ophidianus presented two TTX analogues with m/z 290 and different retention times (3.1 min and 4.6 min). The fragmentation mass pattern of all of them showed a loss of water and m/z 176 and m/z 162 product ions, as TTX. When these samples were analyzed using HRMS, a high error was observed compared to the 11-norTTX theoretical mass. Therefore, it can be concluded that these two TTX analogues have other elemental compositions different from 11-norTTX. Being aware of the error, and since it is in the same species, it could be due to matrix composition and the feeding habits of this particular species [32]. Further studies are needed.

Moroccan Coast
Thirty-eight samples were harvested along the northwestern Moroccan Coast for TTXs, during July 2013 at the 6 sampling points (Figure 1), with two positive hits for TTXs (Table 5).
Though the values detected were all below the reference value implemented in the Netherlands [20], all positives were in gastropods and first report of TTX in these two species, A. depilans and O. celtica, as TTX vectors. O. celtica had quantifiable amounts of TTX, 4-epiTTX and other TTX epimer, and A. depilans was positive below the limit of quantification for TTX and 4-epiTTX.
TTX and 4-epiTTX (m/z 320) were identified in O. celtica (474) and A. depilans (476) samples (Table 5). Besides, in sample 474 another molecule with m/z 320 and different retention time was detected. In this case, the fragmentation pattern was studied and the loss of water and m/z 178 and m/z 162 as product ions were obtained. This TTX analogue was analyzed by HRMS but due to the small quantity it could not be confirmed. As shown in Table 5, TTX and 4-epiTTX were confirmed by HRMS (error −0.6 mDa and 1.7 mDa). The low quantity of 476 sample prevents the confirmation of these compounds by HRMS.

Statistical Analysis
The TTX content per sample is a continuous variable with a large number of samples showing a value of 0 and potentially, overdispersion. Without overdispersion, the best models would be a zero-inflated or not zero-inflated model with Poisson distribution of model error. With overdispersion, we would need to change the error distribution of the model to quasipoisson or negative binomial. Then, in order to decide which statistical model to use, we first tested for overdispersion of our data.
Overdispersion was tested by fitting a generalized function overdispersion test, from the AER R package. This function was applied to a model with Poisson error distribution. In this model, the TTX content was the dependent variable, and the sampling site a factor with 4 levels (São Miguel island, Madeira, continental Portugal and Morocco). TTX content was rounded to appear as a discrete variable (needed for a model with Poisson error distribution). The overdispersion test results were non-significant (z = 1.22, p = 0.11), so it was possible to keep the Poisson model instead of using another one accounting for overdispersion. The goodness of fit of the model error to a Poisson distribution was also tested, with non-significant results (p = 1). Then, it was also not necessary to account for zero-inflation in our data.
An analysis of deviance applied to the previous model showed that 'sampling site' had a significant effect (χ 2 = 882507, p < 0.001). The model coefficients are shown in Table 6, indicating that samples from São Miguel Island had the highest content of TTX, followed by Morocco, whereas Madeira and continental Portugal had the same coefficient value. Due to the large error of the estimate, these latter two coefficients were not significantly different from 0.

Conclusions
In summary, the primary aim of this work was to search for new vectors for TTXs off the Portuguese coast, islands and the northwestern coast of Morocco using UHPLC-HRMS and UHPLC-MS/MS techniques. We report for the first time seven new TTX bearers, 57% of them in gastropods (P. depressa, N. lapillus, O. celtica and A. depilans), 29% in echinoderms (E. esculentus and O. ophidianus) and 14% in fish (S. marmoratus). All the measurable concentrations ranged from 5 to 352,886 TTX eq µg/Kg of SM.
Also, an unknown TTX analogue was detected only in the velvet starfish O. ophidianus, feeding habits can be responsible, but further studies are needed for confirmation.
Regarding geographical tendencies, S. Miguel Island in the Azores was the location with greatest propensity to find these biotoxin groups, followed by the Moroccan Coast; whereas Madeira and continental Portugal had the same coefficient value (≈0). Despite Madeira Island and Morocco having the same latitude, and the sampling being performed one year apart, the tendency values are quite different, which could be due to less anthropogenical interference in Madeira.
We hope this work represents a step forward in the emergent toxin phenomenon and in particularly in the tetrodotoxin theme in line with EFSA's recommendations, stressing the importance of these works for better understanding this emergent subject and protection of public health.
(5 min, 100 Hz) (RK100H, Bandelin SONOREX). A double extraction was performed, extracts were centrifuged at 4495× g for 15 min at 4 • C (Centrifugal-Legend RT, Sorvall, Kendro Laboratory Products, Asheville, NC, USA), supernatants were combined and adjusted to a final volume of 7 mL. Then 1 mL of the extract was cleaned through a C18 solid-phase extraction (SPE), using cartridges (500 mg/3 mL volume from Supelco, Bellefonte, PA, USA) previously conditioned with 6 mL of methanol, followed by 6 mL of water (milliQ). The sample was eluted with 10 mL of 100% methanol and diluted with the same solvent to a final volume of 12 mL. Finally, each sample was concentrated by drying and re-suspending in 1 mL of methanol, and 500 µL were filtered through 0.22 µm filters (UltraFree-MC centrifugal devices, Millipore, Spain) before UHPLC-MS/MS analysis.

UHPLC Conditions
Chromatographic separation was carried out using a 1290 Infinity ultra-high-performance liquid chromatography system coupled to a 6460 Triple Quadrupole mass spectrometer (Agilent Technologies, Waldbronn, Germany). The toxins were separated using a column AQUITY UPLC BEH Amide (2.1 × 100 mm, 1.7 µm, Waters) at 35 • C. Mobile phase A was 100% water with 10 mM formic acid and 10 mM ammonium formate. Mobile phase B was acetonitrile-water (95:5), containing 5 mM formic acid and 2 mM ammonium formate. The gradient program with a flow rate of 0.4 mL/min was started with 100% B and then a linear gradient to 65% B in 7 min. After an isocratic hold time linear of 2 min at 65% B and return to the starting conditions of 100% B in 0.5 min. Finally, 100% B was kept for 3.5 min before the next injection. The samples in the autosampler were cooled to 4 • C and injection volume was 5 µL.

Source Conditions
MS/MS detection was performed using an Agilent G6460C Triple Quadrupole mass spectrometer with an Agilent Jet Stream Electrospray source (Agilent Technologies, Waldbronn, Germany). Source conditions were optimized to achieve the best sensitivity for all compounds drying gas temperature of 250 • C and drying gas flow 5 L/min, nebulizer gas pressure 55 psi (Nitrocraft NCLC/MS from Air Liquid), sheath gas temperature 400 • C and sheath gas flow 12 L/min. The capillary voltage was set to 3000 V in positive mode with a nozzle voltage of 0 V. The fragmentor was 152 and the cell accelerator voltage was 2 for each TTX analogue in this method.
HRMS was acquired by IT-TOF mass spectrometer instrument from Shimadzu (Kyoto, Japan). The mass spectrometer was operated with the following conditions: a heat block and curve desolvation line temperature, 200 • C; nebulizing gas flow, 1.5 L/min; detector voltage, 1.65 kV; and drying gas pressure, 105 kPa. The mass range was calibrated prior to data acquisition employing a direct infusion of IT-TOF standard sample.
The nitrogen generator was a Nitrocraft NCLC/MC from Air Liquid (Meicende, Spain).

UHPLC-MS/MS Analysis: MRM Mode
The Agilent G6460C Triple Quadrupole mass spectrometer was operated in MRM in positive mode and the collision energy (CE) was optimized using MassHunter Optimizer software (Table 8). Two product ions were analyzed per compound, one for quantification and another for confirmation.

UHPLC-MS/MS Analysis: PIS Mode
To confirm and identify TTX analogues, the fragmentation pathway for each molecule was used. The Agilent G6460C Triple Quadrupole mass spectrometer was operated in PIS in positive mode for each TTX analogue with a scan range from 50 to 350 m/z, scan time 120 msec and collision energy from 20 to 50 eV. The characteristic ions were always formed [M+H-H 2