APRIL 2003 VOLUME 16, NUMBER 4 © Copyright 2003 by the American Chemical Society
Articles A Fluorimetric Microplate Assay for Detection and Quantitation of Toxins Causing Paralytic Shellfish Poisoning Maria Carmen Louzao,† Mercedes Rodriguez Vieytes,‡ Ana Garcia Cabado,§ Juan Manuel Vieites Baptista de Sousa,§ and Luis Miguel Botana*,† Departamento de Farmacologia, Facultad de Veterinaria de Lugo, Universidad de Santiago de Compostela, 27002 Lugo, Spain, Departamento de Fisiologı´a, Facultad de Veterinaria de Lugo, Universidad de Santiago de Compostela, 27002 Lugo, Spain, and ANFACO-CECOPESCA, Campus Universitario de Vigo, 36310 Vigo, Pontevedra Received June 28, 2002
Paralytic shellfish poisoning is one of the most severe forms of food poisoning. The toxins responsible for this type of poisoning are metabolic products of dinoflagellates, which block neuronal transmission by binding to the voltage-gated Na+ channel. Accumulation of paralytic toxins in shellfish is an unpredictable phenomenon that necessitates the implementation of a widespread and thorough monitoring program for mollusk toxicity. All of these programs require periodical collection and analysis of a wide range of shellfish. Therefore, development of accurate analytical protocols for the rapid determination of toxicity levels would streamline this process. Our laboratory has developed a fluorimetric microplate bioassay that rapidly and specifically determines the presence of paralytic shellfish toxins in many seafood samples. This method is based on the pharmacological activity of toxins and involves several steps: (i) Incubation of excitable cells in 96 well microtiter plates with the fluorescent dye, bis-oxonol, the distribution of which across the membrane is potential-dependent. (ii) Cell depolarization with veratridine, a sodium channel-activating toxin. (iii) Dose-dependent inhibition of depolarization with saxitoxin or natural samples containing paralytic shellfish toxins. Measuring toxin-induced changes in membrane potential allowed for quantification and estimation of the toxic potency of the samples. This new approach offers significant advantages over classical methods and can be easily automated.
Introduction During red tides (microalgal blooms), filter-feeding shellfish accumulate toxic natural compounds produced * To whom correspondence should be addressed. Tel and Fax: 34 982 252 242. E-mail:
[email protected]. † Departamento de Farmacologia, Facultad de Veterinaria de Lugo Universidad de Santiago de Compostela. ‡ Departamento de Fisiologı´a Animal, Facultad de Veterinaria de Lugo Universidad de Santiago de Compostela. § ANFACO-CECOPESCA, Campus Universitario de Vigo.
by dinoflagellates with no change in their morphology (1). A significant number of these marine toxins exert their effects through interaction with voltage sensitive sodium channels in excitable membranes. In this case, ciguatoxins and brevetoxins perturb normal membrane properties of cells by activating sodium channels (2, 3). In contrast, STX1 and related compounds block ion transport at the voltage-gated Na+ channel and prevent nerve cells from producing action potentials (4, 5). This last action is common to all paralytic shellfish toxins even
10.1021/tx025574r CCC: $25.00 © 2003 American Chemical Society Published on Web 03/01/2003
434
Chem. Res. Toxicol., Vol. 16, No. 4, 2003
though each toxin exhibits a variable relative abundance in natural samples and binds its biological receptor (sodium channel) with different affinity proportional to its intrinsic toxic potency (6, 7). STX is the best known component of paralytic shellfish toxins. However, this group also includes GTX, C toxins, and related compounds such as NeoSTX and dcSTX. Ingestion of shellfish containing high levels of these toxins results in PSP (8). This is a serious health risk and economic challenge to the shellfish industry worldwide. Geographic locations can be toxin free for years, but after dinoflagellate proliferation, toxic shellfish can appear. Most countries have monitoring programs for mollusk toxicity. These monitoring studies involve intensive analysis of a large number of shellfish samples, where the regulatory limit threshold is 80 µg STX equivalent/100 g of mollusk meat. When toxicity is above the limit, the geographical location is closed to shellfish harvesting. Monitoring processes continue until the toxicity drops, at which point harvesting can resume. The standard screening method for marine toxins is the mouse bioassay, which has provided a fairly reliable assessment of risk (9). This is a costly, nonspecific assay with high variability, low sensitivity, and a limited sample throughput. It also involves the use of live animals (10-12). All of those factors have driven the effort to develop other more socially acceptable monitoring programs for PSP. A wide variety of alternative techniques have been described, ranging from instrument-based analytical techniques (13-16) to biological assays (11, 17-21). There has been an increasing interest in the use of biological probes in living cells for better estimation of sample toxicity. We have previously described a solid phase radio receptor assay (22) and a fluorimetric screening (23) for PSP toxins. Here, we report a rapid toxicity test based on a fluorimetric determination of changes in membrane potential induced by PSP toxins through the application of microplate technology for increasing assay efficiency.
Experimental Procedures Materials. Veratridine and gramicidin was from Sigma (Spain). Bis-oxonol was from Molecular Probes (Oregon, U.S.A.). All other chemicals were reagent grade and purchased from Sigma (Spain) and Merck (Germany). Standards of STX were obtained in our laboratory as previously reported (6). The identification of each toxin was performed by HPLC using a C8 column with postcolumn derivatization (6). White solid and white clear bottom microtiter plates (Costar) were used for fluorescence assays. The solution used in the fluorimetric microplate assay contained (mM) as follows: NaCl, 137; KCl, 5; CaCl2, 1; MgCl2, 1.2; KH2PO4, 0.44; NaHCO3, 4.2; glucose, 10, pH 7.4 Methods. (A) Bivalve Samples. Analysis of paralytic shellfish toxins was carried out using mussel samples collected in different culture areas in the Northwest of Spain during PSPproducing dinoflagellate blooms. Mussel (100 g) from each sample was homogenized with HCl 0.18 N (pH 2-4), boiled for 5 min, and filtered. Aliquots (1 mL) of the obtained extract were used for the biological assay. Blank samples (noncontaminated mussels) were randomly selected from routine analysis carried out in our laboratory. 1 Abbreviations: AOAC, Association of Analytical Communities; AU, arbitrary units; bis-oxonol, bis-(1,3-diethylthiobarbituric acid)trimethine oxonol; dcSTX, decarbamoylsaxitoxin; EMEM, Eagle’s minimum essential medium; GTX, gonyautoxins; ND, not detectable; NeoSTX, neosaxitoxin; PSP, paralytic shellfish poisoning; STX, saxitoxin.
Louzao et al. (B) Mouse Bioassay. Following the extraction procedure, the mouse bioassay was performed according to Yasumoto et al. (24). Aliquots (1 mL) of the extract were ip injected into three male mice and observed for 60 min. The test was considered negative when mice survived after an observation period of 60 min. On the contrary, if any mouse died within this time, the test was considered positive. If the mouse died very fast (because of the high concentration of toxin), the sample was diluted with HCl 0.18 N until three mice died between 5 and 7 min. Once the death time for three mice was obtained, Sommers table was used to calculate mouse units (MU) and extrapolate the amount of PSP toxin (STX equivalent/100 g sample). (C) Excitable Cells. Human neuroblastoma cells (BE(2)M17) were grown in EMEM:Hams’s F12 (1:1) supplemented with extra glutamine, nonessential amino acids, 15% fetal bovine serum, and 50 mg/mL gentamicine and 50 µg/mL amphotericin B. Cells were grown in 25 cm2 tissue culture flasks at 37 °C in 5% CO2 and were subcultured by transferring cells released by the application of 0.1% trypsin. (D) Fluorescence Measurement. All reactions were performed in 96 well microtiter plates in a total volume of 200 µL. Changes in membrane potential in BE(2)-M17 cells exposed to PSP toxins were evaluated using the fluorescent probe bisoxonol. Bis-oxonol was added directly to each culture and incubated for 10 min. Fluorescence was measured using a FL600 microplate fluorescence reader (Bio-Tek Instruments, Inc., Vermont, U.S.A.) at the wavelengths 530 nm (excitation) and 590 nm (emission). Depolarization of the cells was induced by the addition of veratridine (40 µM), a sodium channel-activating toxin that enhances sodium ion flux to the cells. Inhibition of neuroblastoma cell depolarization was obtained by the addition of STX or samples containing PSP toxins. At the end of each experiment, 10 µg/mL gramicidin was added to induce complete cell depolarization. Finally, percentage of inhibition was related to STX equivalent concentration. (E) Analysis of Naturally Contaminated Samples with Fluorescent Method and Mouse Bioassay. Naturally contaminated samples of mussel hepatopancreas were analyzed. Results obtained from the new fluorescent method were checked against results obtained from the mouse bioassay. The only required manipulation of extracts to be tested within the fluorescent method was their dilution. (F) Determination by Fluorescent Method of Samples Spiked with Toxin. Analysis of toxins was by the fluorescent method of samples of mussel hepatopancreas spiked with STX. Samples were extracted and analyzed after they were spiked with five concentrations of STX. (G) Statistical Analysis. The experiments were carried out at least four times in duplicate. Results were analyzed using the Student’s t-test for paired data. A probability level of 0.05 or smaller was used for statistical significance. Results are expressed as the mean ( SD.
Results We have previously demonstrated the effectiveness of the optical probe bis-oxonol to detect toxin-induced changes in membrane potential (23). On the basis of those data, we set up a microplate assay to quantify PSP toxins. In this case, it is fundamental to choose the appropriate time points for addition of toxins or measurement of inhibition of depolarization. Assay Characterization. The attached neuroblastoma cells were trypsinized after reaching optimum confluence. The suspension of cells was dispensed (60 000 cells/well) in 96 well plates and incubated overnight (37 °C/5% CO2). The culture medium was removed, and neuroblastoma cells were incubated for 10 min in the assay solution with 4 nM bis-oxonol, whose distribution across the membrane is potential-dependent. After that, fluorescence was measured in a microplate fluorescence
Quantitation of Paralytic Shellfish Toxins
Figure 1. Time course of changes in fluorescence of bis-oxonol in human neuroblastoma cells. Stabilization of the baseline was observed 20-30 min after dye incubation. Stimulation with 40 µM veratridine depolarized neuroblastoma cells and increased their fluorescence. Further addition of 800 ng/mL STX selectively reduced veratridine-induced depolarization, showing decreased bis-oxonol fluorescence. At the end of each experiment, 10 µg/mL gramicidin was added to induce complete cell depolarization. Results are expressed in AU of fluorescence ( SD (N ) 5) (AR ) arbitrary units).
reader set at 530 nm (excitation) and 590 nm (emission). As seen in Figure 1, we registered changes in fluorescence at different times and found stabilization of the baseline in 20-30 min after dye incubation. Veratridine (40 µM) depolarizes neuroblastoma cells and thus increases their fluorescence. Further addition of 800 ng/mL STX selectively reduces veratridine-induced depolarization, showing a decrease of bis-oxonol fluorescence (Figure 1). Finally, complete cell depolarization was obtained with 10 µg/mL gramicidin. On the basis of those data and subsequent experiments, fluorescence of the cells was measured in the microplate reader 30 min after incubation with bis-oxonol to ensure baseline stabilization (basal). The change in fluorescence induced by veratridine was detected 6 min after the addition of the toxin. The PSP effect was measured after 5 min incubation with those toxins. Complete cell depolarization was reached 10 min after gramicidin was added. As can be seen in Figure 2, increasing the amount of cells incubated overnight in each well caused a concomitant increase in fluorescence when neuroblastomas are incubated with 4 nM bis-oxonol. The adequate amount of cells to use is the minimal number that produces a significant response to veratridine or gramicidin. From data presented in Figure 2, 30 000 cells/well for incubation overnight was chosen for assay characterization and use. Concentration of the fluorescent dye for the optimal staining was determined in experiments with concentrations ranging from 2 to 50 nM (Figure 3). We were looking for the concentration of dye that was sensitive to the stimuli that made changes in membrane potential. The adequate dye concentration to use was 4 nM, which was selected for the experiments where we checked PSP toxins. Figure 4 shows changes in fluorescence obtained when different concentrations of STX, which selectively reduced veratridine-induced depolarization, were added to the system. There was a clear relationship between inhibition of depolarization expressed in percentage and toxin concentration represented in a semilogarithmic scale (Figure 5). An important advantage is that experiments could be performed in cellular suspensions. In this case, after the attached neuroblastoma cells were trypsinized, the sus-
Chem. Res. Toxicol., Vol. 16, No. 4, 2003 435
Figure 2. Change in fluorescence of bis-oxonol in human neuroblastoma cells depending on the amount of cells incubated overnight in each well. Fluorescence of the cells was measured in the microplate reader 30 min after incubation with bis-oxonol to ensure baseline stabilization (no toxin). Change in fluorescence induced by 40 µM veratridine was detected 6 min after the addition of the toxin (veratridine). Complete cell depolarization was reached 10 min after 10 µg/mL gramicidin was added (gramicidin). Results are expressed in AU of fluorescence ( SD (N ) 4).
Figure 3. Determination of fluorescent dye concentration for optimal staining. Fluorescence of the cells was measured in the microplate reader 30 min after incubation with different concentrations of bis-oxonol (no toxin). Depolarization was induced with 40 µM veratridine, and fluorescence was detected 6 min after the addition of the toxin (veratridine). At the end of each experiment, 10 µg/mL gramicidin was added to induce complete cell depolarization (gramicidin). Results are expressed in AU of fluorescence ( SD (N ) 4).
pension of cells was washed with the assay solution and directly dispensed (30 000 cells/well) in 96 well plates. Neuroblastoma cells in suspension were incubated for 10 min with bis-oxonol, and fluorescence was measured as previously described (data not shown). Accuracy. To demonstrate the accurancy of the method, nontoxic hepatopancreas from mussels was spiked with known amounts of STX and analyzed in triplicate at every toxic level from 0.016 to 160 µg STX/g of mollusk meat. Results are shown in Figure 6. Figure 7 demonstrates that the relationship between inhibition of depolarization expressed in percentage and toxin concentration is maintained.
436
Chem. Res. Toxicol., Vol. 16, No. 4, 2003
Figure 4. Changes in fluorescence obtained by different amounts of STX. Fluorescence of the cells was measured in the microplate reader after 6 min of incubation with 40 µM veratridine (veratridine). The STX effect was measured after 5 min of incubation with the indicated toxin concentration (STX). In control, a solution where STX is dissolved was added. Results are expressed in AU of fluorescence ( SD (N ) 5).
Louzao et al.
Figure 6. Changes in fluorescence obtained by the addition of extracts of nontoxic hepatopancreas spiked with known amounts of STX. Fluorescence of the cells was measured in the microplate reader after 6 min of incubation with 40 µM veratridine. The STX effect was measured after 5 min of incubation with a sample of the extract containing the indicated toxin concentration (STX). In control, a sample of hepatopancreas extract without toxin was added. Results are expressed in AU of fluorescence ( SD (N ) 5).
Figure 5. Relationship between percentage of inhibition veratridine-induced depolarization by STX and toxin concentration expressed in a semilogarithmic scale.
Screening Results. To check the assay, extracts of naturally PSP-contaminated mussels were added to microplates with neuroblastoma cells loaded with bisoxonol and depolarized with veratridine. The fluorimetric assay included toxin standards with each run obtained from mussels spiked with known amounts of STX. The resultant regression equation from Figure 7 was used to obtain concentration values for unknowns. Using this method, percentage inhibition must fall between 20 and 80% in order to provide accurate results. Samples, which caused >80% inhibitions, were diluted appropriately to fall within this range. Whole extract toxicity expressed as STX equivalents was measured by two methods: mouse bioassay and fluorescent microplate cell assay. Values obtained from individual mussel extracts in the fluorescent microplate cell assay were compared with those obtained from the mouse bioassay (Table 1). Intercomparison. Correlation of dual results of the samples is shown in the plot of Figure 8. Samples not detected by bioassay were rejected for the correlation. Screening results and this intercomparison demonstrated that fluorimetric microplate assay is a suitable test for PSP toxicity detection.
Figure 7. Relationship between percentages of inhibition veratridine-induced depolarization by extracts of nontoxic hepatopancreas spiked with known amounts of STX and toxin concentration expressed in a semilogarithmic scale.
Discussion Shellfish toxicity monitoring is routinely carried out by the mouse bioassay procedure defined by the AOAC (24). However, there is a mounting pressure to develop alternative screening tests for PSP. Biological approaches such as radioligand binding assays, cytotoxicity methods, or ELISA have been developed (11, 17-19, 22, 23, 25, 26). Some of those assays have many advantages but also inconveniences. Each paralytic shellfish toxin exhibits a different specific potency and a highly variable relative abundance in natural samples. Taking into account this fact, immunoassays have a poor predictive value for
Quantitation of Paralytic Shellfish Toxins Table 1. Results Obtained after the Analysis of Mussel Samples by Mouse Bioassay and the Fluorimetric Methoda sample
bioassay (µg/mL)
fluorimetric (µg/mL)
1 2 3 4 5 6 7 8 9 10 A B C
15.4 ( 2.39 2.18 ( 0.39 1.69 ( 0.48 4.37 ( 1.87 0.72 ( 0.30 1.78 ( 0.48 6.16 ( 1.84 1.52 ( 0.5 4.32 ( 1.18 5.75 ( 2 ND ND ND
15.08 ( 0.78 2.32 ( 0.57 1.58 ( 0.26 2.80 ( 0.94 0.99 ( 0.19 1.66 ( 0.18 4.43 ( 0.67 1.15 ( 0.35 4.23 ( 1.12 7.1 ( 3.15 ND 0.03 ( 0.01 ND
a Results are expressed in micrograms of STX per milliliter ( SD (N ) 6).
Figure 8. Correlation of the results of the analysis of mussel samples with the mouse bioassay and the fluorimetric microplate assay.
human toxicity. Those assays rely on antibody-based detection of toxins but have limited cross-reaction of an antibody with different toxin derivatives (12). Alternatively, receptor-binding assays were originally developed to characterize ligand-receptor interactions (27-29) and were subsequently used as techniques based on detection of radioactive compounds. Recent advances in fluorescent dyes allow direct measurement of responses to given toxins with a new methodology. Functional assays using fluorescent probes show exceptional promise as a means of detecting PSP toxins, as well as other marine phycotoxins (12, 19, 22, 23, 26). Related to that, techniques using living cells provide an effective means to measure toxins, especially in situations where toxins are part of a complex mixture of mussel extracts (17, 19, 23). An important step in the refinement of cell-based assays is to simplify the cellular reactions required to generate the functional response of interest (23, 30, 31). In this paper, we described a microplate assay for the determination of sample toxicity by using living neuroblastoma cells and measuring direct changes in membrane potential, induced by toxins, with the fluorescent probe bis-oxonol.
Chem. Res. Toxicol., Vol. 16, No. 4, 2003 437
The implementation of a wide monitoring program of mollusk toxicity requires evaluation of many samples; in this sense, translating the fluorimetric method into a microtiter plate format improves the screening. Bisoxonol is a slow dye that moves across membranes until it reaches electrochemical equilibrium. Cells stained with bis-oxonol exhibit circumferential ringlike patterns, typical of dominating fluorescence intensity from plasma membranes. Cell hyperpolarization is reflected in an increased intracellular cationic dye concentration whereas decreased accumulation reflects depolarization (32). Therefore, fluorescence intensity decreases upon cell hyperpolarization and increases upon cell depolarization. The fluorimetric microplate assay provides an estimate of the integrated toxic potency, since each toxin present in a sample is bound by the receptor with an affinity proportional to its intrinsic toxic potency. Even though the absolute toxicity of these toxins varies considerably, using this assay, total toxicity of all PSP-contaminated samples can be determined. It also allows for measurement of PSP toxins of moderately toxic extracts (even those that are negative to bioassay as we observed in our results). New or poorly known toxins could also be detected. Any modification to a toxin that alters its relative binding to the receptor and thus its detection by this assay would also compromise its ability to elicit a toxic response. In addition to the higher sample test provided by microtiter plate-based assays, measurement of changes in membrane potential is carried out directly within the microtiter plate in which the assay is performed as it happens with other methods (26, 33). This new approach offers significant advantages in terms of rapidity and accuracy when compared to classical methods. PSP toxins were measurable at