Excreted Thiocyanate Detects Live Reef Fishes Illegally Collected Using Cyanide—A Non-Invasive and Non-Destructive Testing Approach

Cyanide fishing is a method employed to capture marine fish alive on coral reefs. They are shipped to markets for human consumption in Southeast Asia, as well as to supply the marine aquarium trade worldwide. Although several techniques can be used to detect cyanide in reef fish, there is still no testing method that can be used to survey the whole supply chain. Most methods for cyanide detection are time-consuming and require the sacrifice of the sampled fish. Thiocyanate anion (SCN−) is a metabolite produced by the main metabolic pathway for cyanide anion (CN−) detoxification. Our study employed an optical fiber (OF) methodology (analytical time <6 min) to detect SCN− in a non-invasive and non-destructive manner. Our OF methodology is able to detect trace levels (>3.16 µg L−1) of SCN− in seawater. Given that marine fish exposed to cyanide excrete SCN− in the urine, elevated levels of SCN− present in the seawater holding live reef fish indicate that the surveyed specimens were likely exposed to cyanide. In our study, captive-bred clownfish (Amphiprion clarkii) pulse exposed for 60 s to either 12.5 or 25 mg L−1 of CN− excreted up to 6.96±0.03 and 9.84±0.03 µg L−1 of SCN−, respectively, during the 28 days following exposure. No detectable levels of SCN− were recorded in the water holding control organisms not exposed to CN−, or in synthetic seawater lacking fish. While further research is necessary, our methodology can allow a rapid detection of SCN− in the holding water and can be used as a screening tool to indicate if live reef fish were collected with cyanide.


Introduction
Thiocyanate (SCN À ) has been widely used in fabric dyeing, photography and electroplating industry, 1 and it has remarkable hazardous effects on both the environment and human health, as it is known to block the iodine uptake by the thyroid gland. 1,2 The accurate determination of SCN À levels in marine water can also be of paramount importance for the detection of live coral reef fishes collected using an illegal and highly destructive technique-cyanide fishing. 3,4 Live fishes traded for human consumption or marine aquariums which have been exposed to cyanide are known to naturally excrete SCN À as a way of selfdepuration. 5 If SCN À levels recorded in the water holding recently imported fishes exceed the basal values recorded for seawater, this can be considered as evidence that imported fishes were actively excreting SCN À and have been illegally collected with cyanide.
Several analytical methods for the determination of SCN À in various types of samples have been previously reported elsewhere, with molecular spectroscopic methods (including UV-Vis spectrophotometry and spectrofluorimetry) being the most widely applied. 1,6,7 The most used spectrophotometric method for thiocyanate determination is based on the formation of a red complex (l ¼ 480 nm) with ferric ion, Fe(SCN) 2+ , in acid medium. 1 Other methods used for determination of SCN À are electrochemical methods (with ion selective electrodes), 8 gas chromatography with electron capture, 9 capillary electrophoresis, 10 sensors based on fluorescence detection, 11 and liquid chromatography coupled to an UV detector (LC-UV). 12 Most of these methods either require sophisticated operation or take long operation times. Additionally, all the methods referred, with exception of LC-UV, cannot be directly applied to the determination of thiocyanate in seawater. Therefore, there is still need for the development of a simple, fast and reliable method for determination of SCN À in seawater.
Optical fiber (OF) sensors and OF based methodologies have been proven to be of high sensitivity and of easy operation for several analytes. 13

Environmental impact
Thiocyanate (SCN À ) has remarkable hazardous effects on both the environment and human health, as it is known to block the iodine uptake by the thyroid gland. Fishes exposed to cyanide, a highly destructive technique used for coral reef's fish collection, are known to naturally excrete thiocyanate as a way of self-depuration; these organisms are traded live and present a high market value both for human consumption and/or the marine aquarium trade. Therefore, obtaining reliable data on SCN À concentration in seawater becomes an important aim both in environmental sciences and sensing technology. An analytical methodology based on optical fiber detection has been developed for thiocyanate assessment in seawater. This methodology is capable of inexpensive, fast and direct analysis of thiocyanate in seawater samples and therefore could be applied to non-invasive and non-destructive detection of coral reef fishes collected using cyanide fishing. the sensitive material/component, the analytical signal generation and characteristics must be understood as an interaction between several physical and chemical factors. Firstly, the analytical signal depends on the wave guide, the polymeric cladding and the analyte properties. Secondly, it is important to take into account the molecular interactions that occur between the analyte molecules and the polymeric film with consequent changes in the intensity of the light power guided through the OF. 27,28 When the analyte molecules contact with the polymeric film, the optical power intensity will vary proportionally to the amount of the analyte present. Without losing sight of the advantages of OF sensors, the aim of this work was to develop an OF based methodology for detection of thiocyanate in seawater samples. The developed methodology was validated by comparison with a high performance liquid chromatography with UV detector (HPLC-UV) methodology and applied to the determination of thiocyanate in seawater samples.

Experimental Preparation of calibrants for thiocyanate determination
Standard solutions of SCN À (4, 50, 100, 200, 300 and 400 mg L À1 ) have been prepared by dilution of NaSCN, (Sigma, Sintra, Portugal) in artificial seawater. Artificial seawater was prepared using freshwater purified by a reverse osmosis unit and mixed with the synthetic salt Crystal SeaÒ produced by Marine Enterprises InternationalÒ (Baltimore, MD, USA), following the instructions of the manufacturer.

Preparation of standard solutions for comparison of methodologies
In order to test and compare the performance of the proposed LC-OF methodology with HPLC-UV, ten different concentrations of SCN À (10, 70, 100, 130, 190, 220, 250, 310 and 370 mg L À1 ) have been prepared, and determined with both methodologies performing five repeated evaluations for each concentration tested.

OF based methodology
The schematic representation of the experimental apparatus of OF based methodology is shown in Fig. 1a. The injector unit (I) was connected to the pump (P) which was connected to the mobile phase flask (MP). The column (C) (150 mm Â 4.6 mm i. d., C30 column modified with 5% polyethylene glycol (PEG) 20 000) was connected to the injector unit (I) and to the analytical tube (AT). The AT had an internal narrowed region of 0.4 cm diameter and 6.5 cm long having inside the OF that is a monomode optical fiber pigtail, core and cladding diameters of 9 and 125 mm, respectively, integrated into a directional 50:50 Y optical coupler (OC), with an angle FC/PC connector and a super FC/PC on the input ends. The optical fiber was previously uncladded, cleaved to 15 mm of the optical path cord and dipped (the cleaved section) into a PEG solution, by dip-coating technique, resulting in the sensitive component of the OF detection. Tests concerning the stability/durability of the sensitive OF section by optical signal monitoring during 3 months and SEM observation of the sensitized OF section revealed that no significant signal degradation and morphology changes were recorded, respectively. Additionally, the sensitive component (uncladded OF section + polymeric film) of the OF methodology can be replaced, after 3 months of continuous operation, in order to maintain high analytical performance after extensive utilization for longer periods of the analytical sensing system. Besides the coated optical fiber and the analytical tube, the detection component of the OF based methodology was constituted by a laser diode optical source (L, Oz Optics, Ottawa, Canada) and a photodiode detector (PD, Oz Optics, Ottawa, Canada). The laser diode (1 mW) optical source was (a) set at 1550 nm for working wavelength and at continuous waveform (CW) regarding the operational mode frequency, to generate the interrogating signal, and (b) connected to the optical coupler (OC). The OC was also connected to the photodiode detector (P), which measures the intensity of the modulated signal, and to a laptop (PC) with homemade software.
Standards of SCN À (4, 50, 100, 200, 300 and 400 mg L À1 ) and seawater samples (20 mL) were introduced by a micro-syringe (Hamilton, Bonaduz, GR, Switzerland) at the top of the injector (I). After separation on the column the analytes reached the analytical tube (AT), which contains the coated OF, generating an analytical signal (Fig. 1b). The changes in the reflected optical power caused by variations in the OF refractive index have been previously discussed by Silva et al., [13][14][15][16][17][18][19][20][21][22][23] and depend on the coating film (PEG) and analyte properties, as well as the chemical interactions which can take place between these two elements. The mobile phase was 300 mM sodium sulfate and 50 mM sodium chloride and the flow-rate was 1.0 mL min À1 .
Calibration models were built by injection of 20 mL of different concentrations (4, 50, 100, 200, 300 and 400 mg L À1 ) of standard solutions. The concentration of thiocyanate was determined by direct interpolation in the calibration curve within their linear dynamic range, and the detection limits were calculated using y ¼ y B + 3s B , where s B is the standard deviation (SD) of the blank signal estimated as s y/x , the residual SD taken from the calibration line, and y B is the blank signal estimated from the intercept taken also from the calibration line. 29

HPLC-UV based methodology
An HPLC-UV based methodology 12 was used in order to test and compare the performance of the OF methodology. The column used was C30 column (150 mm Â 4.6 mm i.d.) modified with 5% PEG 20 000. The mobile phase was 300 mM sodium sulfate and 50 mM sodium chloride and the flow-rate was 1.0 mL min À1 . Standards of thiocyanate were injected (20 mL) on a HPLC-UV (Jasco PU980, Easton, USA) and the wavelength of UV detection was 220 nm.

Sampling of thiocyanate
Seawater samples were collected in clear screw neck glass vials (Fisherbrand, Fisher Scientific, UK) during the months of June and July 2010 from several locations off Aveiro in the Portuguese Coast (Fig. 2). Collected samples were filtered through a 0.45 mm membrane filter and kept under refrigeration at 4 C before further analytical processing. Table 1 summarizes the main analytical characteristics obtained for thiocyanate with the two tested methodologies. The analytical time (less than 6 min) was the same for both approaches, since the same chromatographic components and operational conditions (such as the mobile phase) were used.

Results and discussion
In terms of retention time, linear range and detection limit, the results achieved were in the same order of magnitude for HPLC-UV and OF based methodologies. Both approaches displayed a retention time of approximately 5 min, a linear range from 4 to 400 mg L À1 , while the detection limits were 4 mg L À1 and 3 mg L À1 for the HPLC-UV and OF based methodologies, respectively.
From the calibration study, it was also found that the analytical performance of the OF based methodology reaches a plateau in terms of maximum capacity of detection for concentration values much over the linear range, 30 particularly for concentrations higher than 500 mg L À1 .
Ten standards with different concentrations of thiocyanate were used to evaluate the performance of the OF methodology versus the HPLC-UV methodology. The statistical analysis 31 of the obtained results suggests: (a) that a linear correlation can be established between the two analytical methodologies for the analysis of thiocyanate, with squared correlation coefficient R 2 of 1 (p < 1.59 Â 10 À33 ); (b) that the results obtained with the two analytical methodologies cannot be statistically differentiated since the regression line has an intercept and a slope not significantly different from 0 (p ¼ 0.944) and 1 (p < 1.59 Â 10 À33 ), respectively; and finally (c) that there is a low dispersion level of the results obtained by the two applied analytical methodologies for thiocyanate analysis supported by an extremely narrow interval at a 95% confidence level.
The error, that is ((found value À expected value)/expected value) Â 100) was lower than 0.4% for the OF methodology, and lower than 0.3% for the HPLC-UV methodology.
Apart from keeping the quality of the analysis of SCN À at the same pattern of other usual methods of analysis, i.e. HPLC-UV, the developed methodology compares advantageously due to: (a)  The compact design (Fig. 1) of the OF based methodology allows the integration of the system components into a portable structure for field application, i.e., for thiocyanate in situ analysis. This structure should include a sampling tube and a minivalve, automatically controlled by software and connected to the pump (P, Fig. 1). An electronic platform/software should be also implemented in order to (a) provide direct readings of the analytical signal or (b) database for measurements storage, and (c) remote data acquisition by integration in a wireless network.
After development, comparison and validation against HPLC-UV methodology, the OF methodology was applied to the analysis of thiocyanate in 15 seawater samples collected off Aveiro in the Portuguese Coast. The results from this analysis are summarized in Table 2.
In Table 2, it can be observed that the higher values of thiocyanate were found near the coast (1 km) decreasing to below the detection limit with increasing the distance to the coast. This fact could be related to anthropogenic activities since at off-shore there are larger water bodies which decrease this effect resulting in a decrease of SCN À levels.
Moreover, the levels of thiocyanate found in the water samples collected on the Portuguese coast were in the same range of levels found for seawater samples collected on the Japanese coast (8.1-15.0 mg L À1 ) by Rong et al. 12 Taking into account the complexity of seawater, and in order to access the applicability of the OF based methodology for the analysis of SCN À in this matrix, the potential interference of iodide (concentration of 90 mg L À1 ) in thiocyanate detection was evaluated. No interference could be detected, as iodide produces a peak in a different time window (around 2 min) of the target analyte (SCN À ), as shown in Fig. 3. Moreover, interferences from other analytes such as iodate were also evaluated and no interference was observable since the peak produced was also in a different time window (at 1.20 min).   with a distance to the coast of 1 km, and (b) for a sample of seawater from Barra with a distance to coast of 1 km with thiocyanate standard addition. Iodine is the peak at 2.20 min and thiocyanate is at 5.02 min.

Conclusions
The OF methodology showed a high analytical performance, mainly in terms of linear range (4-400 mg L À1 ) and detection limits (3 mg L À1 ). It has proven to be adequate for the assessment of thiocyanate in seawater and comparable to HPLC-UV methodology. Simple and fast analysis of thiocyanate in seawater, easiness of use and compact design were also analytical features of the developed OF based methodology. In this way, this new methodology can constitute a platform for inexpensive analysis of thiocyanate in environmental samples without iodide interferences. A good example for a practical application of this new methodology is the non-invasive and non-destructive detection of highly priced coral reef fishes collected with cyanide (by rapidly and accurately tracing abnormally high levels of SCN À in the water holding these organisms, rather than sacrificing fish to detect cyanide levels in their tissues and/or blood).