INTRODUCTION
Brevetoxins and ciguatoxins are two marine biotoxins that are produced by some toxic dinoflagellates. Brevetoxins, the causative agents of neurotoxic shellfish poisoning (NSP), are lipid-soluble cyclic polyethers consisting of 10 to 11 transfused rings. Ten natural brevetoxins have been isolated to date (designated PbTX-1 to PbTX-10) (FAO, 2004), and they are produced by the unarmoured dinoflagellate Karenia brevis (formerly Gymnodinium breve or Ptychodiscus breve). Red tide blooms caused by Karenia brevis in the Gulf of Mexico, the east of Florida, New Zealand and Japan (Baden & Adams, Reference Baden and Adams2000) cause massive fish, marine mammal and bird mortality. Brevetoxins are not toxic to shellfish (FAO, 2004), but either consumption of the brevetoxin-contaminated shellfish, or inhalation of the toxic aerosols formed by the wave action of the toxic bloom would result in human intoxication (FAO, 2004). It is not easy to detect the presence of brevetoxins from the contaminated seafood because they are tasteless and odourless. The heat-stable property of the brevetoxins also makes it difficult to eliminate them from the contaminated food.
Ciguatoxins are major toxins responsible for ciguatera fish poisoning (CFP). Similar to brevetoxins, they are also lipid-soluble polyether compounds but consist of 13 to 14 rings fused by ether linkages into a rigid ladder-like structure (FAO, 2004). Ciguatoxins arise from the oxidative biotransformation in the fish from the precursor gambiertoxins (Lehane & Lewis, Reference Lehane and Lewis2000), which are originally produced by the dinoflagellate Gambierdiscus toxicus (Guzman-Perez & Park, Reference Guzman-Perez and Park2000). Gambierdiscus toxicus also produce other toxins like maitotoxin (Gusovsky et al., Reference Gusovsky, Yasumoto and Daly1989) and gambierol (Inoue et al., Reference Inoue, Hirama, Satake, Sugiyama and Yasumoto2003). Maitotoxin is a water-soluble polyether with a C142 chain, 32 ether rings, 28 hydroxyl groups and 2 sulphate esters (Murata et al., Reference Murata, Naoki, Matsunaga, Satake and Yasumoto1994). Gambierol, on the other hand, is a polycyclic ether toxin, and resembles the chromatographic properties of ciguatoxins. Although gambiertoxins, gambierol and maitotoxins are produced by Gambierdiscus toxicus, their pharmacological effects are completely different.
Both brevetoxins and ciguatoxins, or gambiertoxins, have the same receptor target, i.e. site-5 of the α-subunit of the voltage-sensitive sodium channel (VSSC) (Cestele et al., Reference Cestele, Sampieri, Rochat and Gordon1996; Wang & Wang, Reference Wang and Wang2003). VSSC contains one large α-subunit and one or two smaller β-subunits. The α-subunit is a large transmembrane protein that is organized into four repeated homologous domains (I–IV), each consisting of six putative transmembrane segments (S1–S6). They are responsible for the increase in sodium permeability during the initial rapid rising phase of the action potentials in skeletal muscle, nerve and cardiac cells (Cestele et al., Reference Cestele, Sampieri, Rochat and Gordon1996; Wang & Wang, Reference Wang and Wang2003). Binding of brevetoxins or ciguatoxins to VSSC generates membrane depolarization and inhibits its normal inactivation, which leads to a repetitive firing of action potential in resting cells (Berman & Murray, Reference Berman and Murray1999; Cestele & Catterall, Reference Cestele and Catterall2000). Both brevetoxins and ciguatoxins induce wide ranges of gastrointestinal, neurological and cardiac disorders. In cases of severe intoxication, paralysis, coma and death may occur (Baden & Adams, Reference Baden and Adams2000; Dechraoui et al., Reference Dechraoui, Wacksman and Ramsdell2006). Although the molecular action has not been fully characterized, gambierol produces similar symptoms in mice to the ciguatoxins, so gambierol possibly contributes to the effects of CFP too (Guzman-Perez & Park, Reference Guzman-Perez and Park2000). On the contrary, unlike brevetoxins and ciguatoxins that induce sodium ion influx, maitotoxins elicit calcium ion influx in all the tested cell types and tissues (Escobar et al., Reference Escobar, Salvador, Martinez and Vaca1998; Trevino et al., Reference Trevino, De la Vega-Beltran, Nishigaki, Felix and Darszon2006).
Harmful algal blooms and algal toxins not only kill cultured fish and affect coastal ecosystems, but they can also contaminate imported seafood and create a huge economic problem for aquaculture industries. Various assays have been developed to detect and evaluate the toxicity of brevetoxins and ciguatoxins (Gusovsky et al., Reference Gusovsky, Yasumoto and Daly1989; Baden & Adams, Reference Baden and Adams2000; Guzman-Perez & Park, Reference Guzman-Perez and Park2000; FAO, 2004). The earliest methods exploited were in vivo bioassays. Crude shellfish extracts are given to test animals (mice or fish), and the LD50 was determined (Viviani, Reference Viviani1992; Fernandez & Cembella, Reference Fernandez and Cembella1995; Dickey et al., Reference Dickey, Jester, Granade, Mowdy, Moncreiff, Rebarchik, Robl, Musser and Poli1999; FAO, 2004; Campora et al., Reference Bourdelais, Campbell, Jacocks, Naar, Wright, Carsi, Campora, Hokama and Ebesu2006). However, the results obtained from these methods lack specificity as different animals might show a different response and LD50 (Campora et al., Reference Bourdelais, Campbell, Jacocks, Naar, Wright, Carsi, Campora, Hokama and Ebesu2006). Moreover, large numbers of animals were required for testing a particular concentration of a toxin, which can be problematic, both ethically and financially (FAO, 2004). Cell-based in vitro bioassays, on the other hand, are much easier to handle. Cell cultures are exploited instead of animals, which is an ethically more acceptable approach. Cell cytotoxicity assays, which use either MTT(3-(4,5- Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) or XTT (2,3-bis-(2-methoxy-4-nitro-5-sulphophenyl)-2H- tetrazolium-5-carboxanilide), or reporter gene assays, which utilize luciferase-catalysed light generation (Jellett et al., Reference Jellett, Marks, Stewart, Dorey, Watson-Wright and Lawrence1992; Manger et al., Reference Manger, Leja, Lee, Hungerford and Wekell1993, Reference Manger, Lee, Leja, Hungerford, Hokama, Dickey, Granade, Lewis, Yasumoto and Wekell1995; Dechraoui et al., Reference Dechraoui, Naar, Pauillac and Legrand1999; Guzman-Perez & Park, Reference Guzman-Perez and Park2000; Fairey et al., Reference Fairey, Dechraoui, Sheets and Ramsdell2001; FAO, 2004) have been developed to quantify the brevetoxins and ciguatoxins. When compared to immuno-biochemical methods, e.g. enzyme-linked immunosorbent assay (ELISA) (Van Dolah et al., Reference Van Dolah, Finley, Haynes, Doucette, Moeller and Ramsdell1994; Guzman-Perez & Park, Reference Guzman-Perez and Park2000; Bourdelais et al., Reference Bourdelais, Campbell, Jacocks, Naar, Wright, Carsi, Campora, Hokama and Ebesu2006) and HPLC methods (Baden & Adams, Reference Baden and Adams2000; Guzman-Perez & Park, Reference Guzman-Perez and Park2000; FAO, 2004), fluorimetric microplate assays are relatively rapid, simple and less expensive.
Effective monitoring of harmful algal blooms is generally the only way to tackle these environmentally devastating events. There is also an ever-increasing trend of toxic blooms. Regulatory agencies have to make decisions on aquaculture, fisheries and leisure activities with very limited information. Very often, the same species has both toxic and non-toxic strains; or the same strain is toxic only under certain conditions. There seems to be a lack of effort in developing direct monitoring assays for toxins at the pre-bloom and mix-assemblage situation. The present investigation aims to test the fluorescent signals, as monitored by commonly used cation-binding dyes. Cross-talk between signalling pathways can also cause changes in other intracellular ion concentrations. These characterized patterns of responses can be immediately used for laboratory monitoring of toxic algal cultures, including the possible interactions of different toxic species. Characterization of intracellular sodium and calcium in response to purified toxins and algal extracts will be the first step in designing laboratory assay using multiplex fluorescent-assays (MFAs) for their possible use in routine laboratory monitoring programmes.
MATERIALS AND METHODS
Materials
Pure brevetoxins PbTX-3 were obtained from Calbiochem. The sodium ionophore monensin, the Na+/K+-ATPase inhibitor ouabain, and the calcium ionophore ionomycin were purchased from Sigma. SBFI-AM was from Molecular Probes, Invitrogen. Microplates used for the assays were black clear-bottom 96-well microplates for tissue culture from Costar. Dulbecco's modified Eagle medium (DMEM) and Hank's balanced salt solution (HBSS) were purchased from Gibco, Invitrogen. HBSS contains (mM): 1.26 CaCl2, 0.493 MgCl2 · 6H2O, 0.407 MgSO4 · 7H2O, 5.33 KCl, 0.441 KH2PO4, 4.17 NaHCO3, 137.93 NaCl, 0.338 Na2HPO4, 5.56 D-Glucose. All other chemicals were obtained from Sigma.
Cell culture
DINOFLAGELLATE CELL CULTURE
Dinoflagellate Gambierdiscus toxicus (CCMP 401) and Karenia brevis (CCMP 2281) cells were maintained in K and L medium respectively at 17°C, under photon flux of 50 µmol m−2s from fluorescent tubes (Philips daylight) in 12:12 hours light/dark cycles. Cells in exponential growth phase were maintained in batch cultures (100 ml) by serial transfer regularly. Batch cultures (1–5 × 105 cells/ml) for the preparation of algal extracts were grown in plastic 175 ml tissue culture flasks (Nunc Corporation) or one litre glass flat-bottom flasks for G. toxicus and K. brevis respectively.
MAMMALIAN CELL CULTURE
Neuroblastoma cells NG108-15 and fibroblast 3T3-L1 were maintained with DMEM, 0.1 unit/ml of penicillin, 0.1 mg/ml of streptomycin and 5% of foetal bovine serum (FBS). Cells were grown on 100-mm plastic tissue culture dishes at 37°C in an incubator with 5% CO2 and a humidified atmosphere. Trypsin-detached confluent cells were used for all fluorescent measurement unless mentioned. We have also tested BE(2)-M17 cells, another neuronal cell line, as possible host cells for these assays. However, we found that NG108-15 was generally more sensitive to the effects of the tested algal toxins (data not shown).
Intracellular calcium level measurement
When the cultured cells reached optimum confluency in 100-mm tissue culture dishes, they were detached by trypsin. Approximately 1 × 105 cells were seeded on each well on a microplate and incubated at 37°C overnight. Following overnight incubation, the medium was discarded and replaced with 2 µM Fluo4-AM in HBSS with 0.2% pluronic acid F-127 and 1% FBS for 1 hour at 37°C. Resting fluorescence signals were measured by spectrophotometer GEMINI XS (Molecular Device) (excitations at 488 nm and emission at 520 nm). Ten microlitres of ionomycin (positive control) or PbTX-3 was added and the corresponding changes in fluorescent intensity were recorded. The temperature was maintained at 37°C throughout the whole experiment.
Intracellular sodium level measurement
The sodium-sensitive dye, SBFI, is a selective sodium ion indicator for the fluorimetric determination of sodium ion concentrations. Its acetoxymethyl (AM) ester form is cell permeable, and provides resolution of physiological concentrations of sodium ions (Diarra et al., Reference Diarra, Sheldon and Church2001; Bicalho et al., Reference Bicalho, Guatimosim, Prado, Gomez and Romano-Silva2002; Maaser et al., Reference Maaser, Hopfner, Kap, Sutter, Barthel, von Lampe, Zeitz and Scherubl2002). When the cultured cells reached optimum confluency in 100-mm tissue culture dishes, they were rinsed once with HBSS, and then incubated with 10 µM of SBFI-AM in HBSS with 0.2% pluronic acid F-127 and 1% FBS for 3 hours at 37°C. After dye loading, the cells were rinsed twice with HBSS, detached from the dish and approximately 1 × 105 cells were transferred into each well of a 96-well microplate. The cells were further incubated in HBSS containing 1% FBS and 300 µM ouabain at 37°C for 30 minutes. After the second incubation, the resting fluorescence signals were measured with excitations at 330 nm and 380 nm, and emission at 470 nm. Ten microlitres of monensin (positive control) or PbTX-3 was added and the corresponding changes in fluorescent intensity were recorded. The temperature was maintained at 37°C throughout the whole experiment.
Calcium free HBSS, which contains 5 mM EGTA but not CaCl2 was used throughout the experiments to exclude the effects of extracellular calcium ions. The concentrations of other ions in the buffer used for washing and incubation steps were kept the same as normal HBSS.
Preparation of dinoflagellate cell extracts
KARENIA BREVIS CELL EXTRACTS
In order to investigate the potential application of fluorimetric assays to harmful algal blooms monitoring, it is necessary to carry out preliminary tests with algal extracts and purified toxins. Karenia brevis was harvested by centrifugation (1000 ×g) once the cells had grown up to 5000 cells per ml. Methanol:water (1:1) was added to the cell pellet prior to sonication. The cell debris was separated by centrifugation (4000 ×g), and the supernatant (MeOH/H2O extract) was saved. The cell debris portion was twice subjected to the same extraction procedure. The remaining residue was then extracted with MeOH/H2O/acetone (50:50:0.2) three times, and the supernatant (MeOH/H2O/acetone extract) was saved. Both extract portions were concentrated into powder by freeze-drying. Dimethyl sulphoxide (DMSO) was used as solvent for the assay measurement.
GAMBIERDISCUS TOXICUS CELL EXTRACTS
Gambierdiscus toxicus was harvested by centrifugation (1000 ×g) once the cells had grown up to 200 cells per ml. Methanol was added to the cell pellet, which was then subjected to sonication. The supernatant (cell extract) was collected by centrifugation (4000 ×g). The cell extract was concentrated into powder by freeze-drying. DMSO was used as solvent for the assay measurement.
RESULTS
Calcium and sodium dye calibrations
First of all, the calcium dye, Fluo4-AM, and the sodium dye, SBFI-AM, were calibrated using the calcium and sodium ionophores, ionomycin (10 µM), and monensin (10 µM) respectively (Merritt et al., Reference Merritt, McCarth, Davies and Moores1990; Lamont et al., 1998; Diarra et al., Reference Diarra, Sheldon and Church2001). The baseline fluorescence was stabilized within 30 minutes. An increase in intracellular calcium (Figure 1A, B) and sodium levels (Figure 1C, D) were observed immediately for both NG108-15 and 3T3-L1 cell lines following the addition of their specific ionophore. For intracellular sodium level measurements, additional sodium ions (50 mM) were supplemented together with the monensin in order to amplify the fluorescent signal caused by larger sodium influx.
Fig. 1. Calibration of calcium indicator Fluo4-AM and sodium indicator SBFI-AM. Ionomycin (10 µM) was added to Fluo4-AM-loaded NG108-15 (A) and 3T3-L1 (B) ells. There is a rise in fluorescence intensities, which directly correlates with the intracellular calcium levels, after the addition. Monensin (10 µM) was added to SBFI-AM-loaded NG108-15 (C) and 3T3-L1 (D) cells. After addition of monensin, an increase in fluorescence intensities that is directly related to intracellular sodium level was observed.
PbTX-3 raised the intracellular sodium level but not the calcium level
Different concentrations of PbTX-3 (0.05–5 µM) were added to the SBFI-AM-loaded cells, and the response signals were recorded for 60 minutes. PbTX-3 increased the SBFI-mediated fluorescence of NG108-15 cells (Figure 2A). Immediately after PbTX-3 addition, there was an increase in fluorescence. Ten minutes later, the fluorescent signal gradually decreased for the 0, 0.05 and 0.1 µM PbTX-3 treatments. No difference in the fluorescent levels was observed for 0.5 and 1 µM of PbTX-3 treatments, whereas a gradual increase in fluorescent level was observed for the treatment with 5 µM PbTX-3. The final fluorescent readings for different concentrations of PbTX-3 were compared and a concentration-dependent increase of fluorescence was observed (Figure 2B). The lowest concentration that could be detected (detection limit) for the assay was 0.5 µM PbTX-3. A similar result was also observed for the cell line 3T3-L1 (Figure 2C, D). The results indicated that the assay could detect a sodium influx caused by PbTX-3.
Fig. 2. Effects of PbTX-3 on intracellular sodium and calcium levels of NG108-15 and 3T3-L1 cells. (A) Response of NG108-15 cells to PbTX-3. Upon the addition of PbTX-3, an increase in SBFI-mediated fluorescent level was observed; (B) a bar chart of fluorescence versus PbTX-3 concentrations in NG108-15 cells. The changes in fluorescence were concentration-dependent; (C) response of 3T3-L1 cells to PbTX-3; (D) a bar chart of fluorescence versus PbTX-3 concentrations in 3T3-L1 cells; (E) Fluo-4-mediated fluorescent level of NG108-15 cells in response to PbTX-3. There is no observable intracellular calcium increase after the addition of the brevetoxins.
Although a small increase in fluorescent signal was observed for all concentrations of PbTX-3 in the calcium assay, the increase in fluorescent level for all concentrations tested was the same as the control (0 µM of PbTX-3) (Figure 2E). Unlike causing an intracellular sodium increase, PbTX-3 did not induce calcium influx in NG108-15 cells.
Karenia brevis cell extracts do not affect either intracellular calcium or sodium level
The effects of the Karenia brevis extracts were also monitored by using the assays mentioned above. Cell extracts corresponding to the cell densities 100, 500 and 1000 cells were tested. Surprisingly, both extract fractions (MeOH/H2O and MeOH/H2O/acetone) did not cause any calcium (Figure 3A, B) or sodium influx (Figure 3C, D) in NG108-15 cells.
Fig. 3. Effects of Karenia brevis cell extracts on intracellular sodium and calcium levels of NG108-15 cells (A & B). There is no SBFI-mediated fluorescent level increase after the addition of the cell extracts from either MeOH/H2O (1:1) (A) or MeOH/H2O/acetone (50:50:0.2) (B) fractions. (C & D) No Fluo-4-mediated fluorescent increase was observed after the addition of the cell extracts from either MeOH/H2O (1:1) (C) or MeOH/H2O/acetone (50:50:0.2) fractions.
Gambierdiscus toxicus cell extracts affects both intracellular calcium and sodium level
Morales-Tlalpan & Vaca (Reference Morales-Tlalpan and Vaca2002) reported that maitotoxin could activate both calcium and sodium influx. However, in the absence of extracellular calcium (i.e. in the presence of EGTA), there was no sodium influx in response to maitotoxin. By performing the sodium assay, and the calcium assays with and without external calcium, the activity of the two toxins, sodium mobilizing ciguatoxins and calcium mobilising maitotoxin, from the dinoflagellate Gambierdiscus toxicus could possibly be distinguished. Our results showed a concentration-dependent response towards maitotoxin (Figure 4A), which was the same as that in Morales-Tlalpan & Vaca (Reference Morales-Tlalpan and Vaca2002). Maitotoxin caused an increase in the intracellular sodium level in NG108-15 cells, but the sodium level decreased in a calcium free (no Ca2+) condition (Figure 4C). For extraction of G. toxicus toxins, methanol, which can dissolve both water and lipid soluble compounds, was used as solvent. The cell extracts caused a cell-density dependent calcium increase in NG108-15 cells (Figure 4B). In response to G. toxicus cell extracts, SBFI-mediated fluorescence increased in the absence of extracellular calcium. The fluorescent level was slightly higher than that in response to maitotoxin. The results indicated that by using one calcium assay and two sodium assays (with or without external calcium) we can distinguish the activity between sodium mobilizing and calcium mobilizing toxins.
Fig. 4. Effects of maitotoxin and Gambierdiscus toxicus cell extracts on both intracellular calcium and sodium levels of NG108-15 cells. (A) A concentration-dependent of Fluo-4-mediated fluorescent increase was observed in response to maitotoxin; (B) a cell density-dependent increase in fluorescent level was also observed in response to G. toxicus cell extracts; (C) an SBFI-mediated fluorescent level increase was observed in response to 1 nM maitotoxin in the presence of extracellular calcium ions. In calcium free condition (no Ca2+), the level of sodium influx decreased; (D) an intracellular sodium level increase was observed in response to 500 cells of G. toxicus.
DISCUSSION
Many marine toxins target a specific cation channel. The cell-based fluorimetric assays employed used cation-specific dyes to detect the variation in intracellular cation levels caused by the molecular action of the toxins. Two types of cell permeable ion dye, Fluo4-AM and SBFI-AM, were exploited for the detection of calcium and sodium influx respectively. Our results showed that PbTX-3 caused a concentration-dependent sodium influx, but not a calcium influx, in NG108 cells. An increase of SBFI-AM mediated fluorescence was also observed in the control, but declined over time. This was probably due to the pumping of sodium ions out of the cells. In PbTX-3-treated cells, the level of SBFI-mediated fluorescence remained high, probably attributed to equalization of efflux and influx of sodium ions.
The cell extracts from K. brevis, the principal dinoflagellate that produces brevetoxins, could not induce sodium influx in our experiments. One possible explanation is the presence of an antagonist in the cell extracts which can reverse the action of brevetoxins. Bourdelais et al. (Reference Bourdelais, Campbell, Jacocks, Naar, Wright, Carsi and Baden2004) reported that a short natural polyether product called brevenal was isolated from K. brevis. Brevenal could compete with brevetoxin for the same site 5 of voltage-sensitive sodium channel in their experiments. It is a natural inhibitor of brevetoxin action in sodium channel receptors (Bourdelais et al., Reference Bourdelais, Campbell, Jacocks, Naar, Wright, Carsi, Campora, Hokama and Ebesu2004). Our data suggest a separation of brevetoxin from brevenal may be required for the monitoring the toxic effects of the toxins.
Gambierdiscus toxicus extracts were able to induce both SBFI-AM-mediated and Fluo-4-mediated fluorescence. Maitotoxin alone was able to induce both calcium and sodium influx (Morales-Tlalpan & Vaca, Reference Morales-Tlalpan and Vaca2002; Figure 4C). However, a significant increase of SBFI-AM-mediated fluorescence was observed, even in the absence of extracellular calcium. This confirms that our extract contained both maitotoxin and gambiertoxin. These data also support the use of a duplex assay, in the presence and absence of extracellular calcium, to monitor the two toxins.
The sensitivity of fluorimetric assays to low cell number of toxic algae point to their possible use in laboratory monitoring for toxin production of toxic algae. Their simplicity also supports the assays as an alternative to more equipment-based analytical methods. Most known algal toxins are modulators of ion channels, with the ability to affect intracellular calcium and sodium level. As previously reported by other groups (Louzao et al., Reference Louzao, Cagide, Vieytes, Sasaki, Fuwa, Yasumoto and Botana2006), we were also able to observe dose-dependent fluorescent response of membrane potential dye in response to saxitoxins (data not shown). We have also carried out preliminary tests of the background level of potential channel modulating agents from seawater at Port Shelter (off the Hong Kong University Pier). Using methanol to extract algal filtrate (>1 µm, <10 µm ) from one litre of seawater, we could not detect any background level of fluorescent-inducing agents throughout the year. However, addition of 50–100 toxic cells (G. toxicus) to the same seawater was able generate a positive signal (data not shown). These data point to the potential use of MFA as part of a routine monitoring protocol in pre-bloom condition. During toxic algal bloom, an algal extract from a fixed assemblage of phytoplankton may also generate specific signal that can lead to the delineation of toxic agents. It is possible that a multiplex of fluorimetric assays (MFA), using Ca-binding, Na-binding and membrane potential dyes will provide the laboratory-based assays that potentially complement existing species identification-based monitoring programmes. However, further works are required to characterize MFA responses to more known toxins.
ACKNOWLEDGEMENTS
The present project was partly supported by an Environment and Conservation Fund grant (ECWW03/04.SC02) to J.T.Y.W. and Y.H.W. We wish to thank Mike Bennett for reading the manuscript.
