Assessment of Specific Binding Proteins Suitable for the Detection of

Jun 21, 2007 - Scott A. WalperGuillermo Lasarte AragonésKim E. SapsfordCarl W. ..... Gary R. McCoy , Sara McNamee , Katrina Campbell , Christopher T...
0 downloads 0 Views 165KB Size
Download PDF
Anal. Chem. 2007, 79, 5906-5914

Assessment of Specific Binding Proteins Suitable for the Detection of Paralytic Shellfish Poisons Using Optical Biosensor Technology Katrina Campbell,*,† Linda D. Stewart,† Gregory J. Doucette,‡ Terence L. Fodey,†,§ Simon A. Haughey,| Natalia Vilarin˜o,⊥ Kentaro Kawatsu,# and Christopher T. Elliott†

Institute of Agri-Food and Land Use (IAFLU), Queen’s University, David Keir Building, Stranmillis Road, Belfast, Northern Ireland, United Kingdom, BT9 5AG, Marine Biotoxins Program, NOAA/National Ocean Service, Center for Coastal Environmental Health and Biomolecular Research, 219 Fort Johnston Road, Charleston, South Carolina 29412, Agri-Food and Biosciences Institute (AFBI), Stoney Road, Stormont, Belfast, Northern Ireland, United Kingdom, BT4 3SD, Xenosense Ltd., Unit 18 The Innovation Centre, The Northern Ireland Science Park, Queen’s Road, Queen’s Island, Belfast, Northern Ireland, BT3 9DT, Departamento de Farmacologı´a1, Facultad de Veterinaria, Universidad de Santiago de Compostela, Campus Universitario, 27002 Lugo, Spain, and Division of Bacteriology, Osaka Prefectural Institute of Public Health, 3-69, Nakamichi 1-chome, Higashinari-ku, Osaka 537-0025, Japan

Paralytic shellfish poisoning (PSP) toxin monitoring in shellfish is currently performed using the internationally accredited AOAC mouse bioassay. Due to ethical and performance-related issues associated with this bioassay, the European Commission has recently published directives extending procedures that may be used for official PSP control. The feasibility of using a surface plasmon resonance optical biosensor to detect PSP toxins in shellfish tissue below regulatory levels was examined. Three different PSP toxin protein binders were investigated: a sodium channel receptor (SCR) preparation derived from rat brains, a monoclonal antibody (GT13A) raised to gonyautoxin 2/3, and a rabbit polyclonal antibody (R895) raised to saxitoxin (STX). Inhibition assay formats were used throughout. Immobilization of STX to the biosensor chip surface was achieved via aminocoupling. Specific binding and inhibition of binding to this surface was achieved using all proteins tested. For STX calibration curves, 0-1000 ng/mL, IC50 values for each binder were as follows: SCR 8.11 ng/mL; GT13-A 5.77 ng/mL; and R895 1.56 ng/mL. Each binder demonstrated a different cross-reactivity profile against a range of STX analogues. R895 delivered a profile that was most likely to detect the widest range of PSP toxins at or below the internationally adopted regulatory limits. Paralytic shellfish poisoning (PSP) is the most prevalent shellfish intoxication syndrome on a worldwide basis.1 The PSP toxins are produced by both prokaryotic and eukaryotic organ* Corresponding author. E-mail: [email protected]. Fax: 0044 (0) 2890976513. † Queen’s University. ‡ NOAA/National Ocean Service. § Agri-Food and Biosciences Institute (AFBI). | Xenosense Ltd. ⊥ Universidad de Santiago de Compostela. # Osaka Prefectural Institute of Public Health. (1) Hallegraeff, G. M. Phycologia 1993, 32, 79-99.

5906 Analytical Chemistry, Vol. 79, No. 15, August 1, 2007

isms, predominantly cyanobacteria and dinoflagellates, respectively.2 The toxins responsible for PSP are a suite of heterocyclic guanidines of which there are over 21 known analogues of the parent compound (Figure 1), saxitoxin (STX). The structure of STX was first described in 1957,3 and all known analogues have been structurally well defined.4,5 The structures of the STX analogues vary by alternating combinations of hydroxyl and sulfate substitutions at the four sites R1-R4 (Figure 1). Based on these substitutions, the toxins can be divided into the following groups in general order of descending toxic potency: the carbamate, decarbamoyl, and sulfocarbamoyl toxins. Substitutions at the R4 site result in the most significant changes to the toxicity of these compounds. A fourth group, the deoxydecarbamoyl toxins, has yet to be purified in sufficient quantities to assess their potency. STX and its analogues are potent neurotoxins that bind with a high affinity to the voltage-dependent sodium channel, preventing channel opening and specifically inhibiting the excitation current in nerve and muscle cells, ultimately resulting in paralysis.6 Their effect on the body illustrates the importance of providing reliable methods for monitoring PSP toxin levels in shellfish. Traditionally, the mouse bioassay, now standardized by the AOAC,7 has been used to quantitatively determine the PSP toxicity of shellfish in many countries. The binding affinity of STX analogues to the sodium channel varies in proportion to their toxicity as measured by mouse ip injection.8,9 However, for the routine analysis of large sample numbers, the mouse bioassay has a low sample throughput and is costly as well as labor intensive. Inaccuracies caused by (2) Shimizu, Y. Annu. Rev. Microbiol. 1996, 50, 431-465. (3) Schantz, E. J.; Mold, J. D.; Stanger, D. W.; Shavel, J.; Riel, F. J.; Bowden, J. P.; Lynch, J. M.; Wyler, R. S.; Riegel, B. R.; Sommer, H. J. Am. Chem. Soc. 1957, 79, 5230-5235. (4) Schantz, E. J. Ann. N. Y. Acad. Sci. 1986, 479, 15-23. (5) Kao, C. Y. Ann. N. Y. Acad. Sci. 1986, 479, 52-67. (6) Kao, C. Y. Pharmacol. Rev. 1966, 18, 997-1049. (7) Hollingworth, T.; Wekell, M.M. In Official Methods of Analysis, 15th ed.; Hellrich, K., Ed.; AOAC: Arlington, VA. 1990; pp 881-882. (8) Doucette, G. J.; Logan, M. L.; Ramsdell, J. S.; Van Dolah, F. M. Toxicon 1997, 35, 625-636. (9) Oshima, Y. J. AOAC Int. 1995, 78, 528-532. 10.1021/ac070342o CCC: $37.00

© 2007 American Chemical Society Published on Web 06/21/2007

Figure 1. Structures of paralytic shellfish poisoning toxins.

high salt concentrations10 and accumulated zinc in the shellfish11 as well as high variability in assay response12-14 are several other limitations of this in vivo method. Moreover, there are substantive ethical issues and consequences of using live mammals for toxicity testing.15 Hence, the development of alternative user-friendly, rapid, and reliable screening assays to replace the AOAC mouse bioassay has become a priority within laboratories testing routinely for PSP toxins. Within Europe, the European Animal Protection Legislation (Council Directive 86/609/EEC) was introduced to ensure progress away from animal experimentation to scientifically acceptable, nonanimal procedures fully validated to an international standard. However, the lack of sufficient amounts of standard reference material for all PSP toxin congeners has greatly inhibited progress. Documented methods of PSP toxin analysis fall into three main categories: bioassays (in vivo and in vitro), biochemical assays, and chemical analyses. Bioassays reported include the mouse bioassay as noted above, sodium channel receptor assays,8 cytotoxicity assays,16,17 and tissue-based biosensors;18 biochemical (10) Schantz, E. J.; McFarren, E. F.; Schafer, M. L.; Lewis, K. H. J. AOAC Int. 1958, 41, 160-168. (11) Aune, T.; Stabell, O. B.; Nordstoga, K.; Tjøtta, K. Nat. Toxins 1998, 7, 141158. (12) Park, D. L.; Adams, W. N.; Graham, S. L.; Jackson, R. C. J. AOAC Int. 1986, 69, 547-550. (13) LeDoux, M.; Hall, S. J. AOAC Int. 2000, 83, 305-310. (14) Earnshaw, A. Marine Toxins, Pilot Study August 2003; Report Food Analysis Performance Assessment Scheme, Central Science Laboratory, Sand Hutton, York, UK, 2003 (15) Jellett, J. F.; Roberts, R. L.; Laycock, M. V.; Quilliam, M. A.; Barrett, R. E. Toxicon 2002, 40, 1407-1425. (16) Jellett, J. F.; Marks, L. J.; Stewart, J. E.; Dorey, M. L.; Watson-Wright, W.; Lawrence, J. F. Toxicon 1992, 30, 1143-1156. (17) Okumura, M.; Tsuzuki, H.; Tomita, B. Toxicon 2005, 46, 93-98.

assays that have been described are immunoassays15,19-21 and electrochemical immunosensors;22 and chemical analyses developed include chromatographic or electrophoretic techniques9,23-26 and mass spectrometry.27,28 The majority of these previously published methods were designed to either replace or reduce the use of the mouse bioassay. However, analytical complications, such as interference by nonspecific matrix effects, excessively complex sample extraction protocols, minimal sensitivity toward certain PSP toxin analogues, or limited availability of standard reference materials, have compromised widespread use of these methods for routine monitoring. The only first action method accepted by the AOAC as an alternative to the MBA is the precolumn derivatization, high-performance liquid chromatogra(18) Cheun, B. S.; Loughran, M.; Hayashi, T.; Nagashima, Y.; Watanabe, E. Toxicon 1998, 36, 1371-1381. (19) Chu, F. S.; Hsu, K. H.; Huang, X.; Barrett, R.; Allison, C. J. Agric. Food Chem. 1996, 44, 4043-4047. (20) Garthwaite, I.; Ross, K. M.; Miles, C. O.; Briggs, L. R.; Towers, N. R.; Borrell, T.; Busby, P. J. AOAC. Int. 2001, 84, 1643-1648. (21) Kawatsu, K.; Hamano, Y.; Sugiyama, A.; Hashizume, K.; NoGuchi, T. J. Food Prot. 2002, 65, 1304-1308. (22) Carter, R. M.; Poli, M. A.; Pesavento, M.; Sibley, D. E. T.; Lubrano, G. J.; Guilbault, G. G. Immunomethods 1993, 3, 128-133. (23) Mosley, S.; Ikawa, M.; Sasner, J. J. Toxicon 1985, 23, 375-381. (24) Thibault, P.; Pleasance, S.; Laycock, M. V. J. Chromatogr. 1991, 542, 483501. (25) Lawrence, J. F.; Niedzwiadek, B. J. AOAC Int. 2001, 84, 1099-1108. (26) Kele, P.; Orbulescu, J.; Gawley, R. E.; Leblanc, R. M. Chem. Commun. 2006, 14, 1494-1496. (27) Quilliam, M.; Hess, P.; Dell’Aversano, C. In Mycotoxins and Phycotoxins in Perspective at the Turn of the Millennium; De Koe, W. J., Samson, R. A., Van Egmond, H. P., Gilbert, J., Sabino, M., Eds.; Proceedings of the X International IUPAC Symposium on Mycotoxins and Phycotoxins: Brazil; 2000; pp 383-391. (28) Dell’ Aversano, C. D.; Hess, P.; Quilliam, M. A. J. Chromatogr., A 2005, 1081, 190-201.

Analytical Chemistry, Vol. 79, No. 15, August 1, 2007

5907

phy-fluorescence detection (HPLC-FLD) method described by Lawrence and co-workers.29 However, initial reports suggest that this method is labor intensive and that excessive quantities of standard reference material are required for producing complete toxin profiles.30 Current opinion thus appears to indicate that development of a convenient, quick, and cost-effective method for detecting all PSP toxins of regulatory relevance remains a critical need for monitoring these compounds. Optical biosensors based on surface plasmon resonance technology are an important tool for biomedical research and are being used increasingly by academic and industrial scientists interested in molecular interactions and drug/toxin analysis in food safety. Given the success of this platform for detecting lowlevel chemical contaminants in foods, these biosensors show a strong potential as an alternative strategy for monitoring PSP toxins. Biacore biosensors have been applied previously for toxin analysis in food products. Assays have, in fact, been developed to detect another marine algal toxin, domoic acid,31,32 as well as staphylococcal enterotoxin B,33-35 and the fungal aflatoxin B1.36 These biosensor-based assays measure the interaction of a specific biological recognition element or binder with a target analyte (e.g., toxin) immobilized onto the sensor chip surface. The biological binders most frequently adopted in assay development for their ability to interact with PSP toxins are sodium channel receptors (SCR) and antibodies raised against various toxin analogues. Sodium channel receptors from crude rat brain membrane preparations have been used in radioligand binding assays37 and more recently configured into a microtiter filtration plate format.38,39 The fact that all STX analogues bind to these receptors with affinities that vary according to their toxicity8 indicates that receptor-based competitive binding assays provide a measure of a sample’s integrated toxic potency, irrespective of which PSP toxins are present. A saxitoxin-specific receptor referred to as saxiphilin, occurring naturally in various amphibians and terrestrial invertebrates, has also been incorporated into a competitive binding assay employing a radioactive end point.40 The first antibodies produced against PSP toxins were reported in 1964;41 however, it was decades later that immunoassays were actually utilized to monitor PSP toxins in shellfish. Several antibody-based approaches for detecting these toxins have been described, including radioimmunoassays42 and various competitive enzyme immunoassays (EIA). Indirect EIAs have been developed (29) Lawrence, J. F.; Niedzwiadek, B.; Menard, C. J. AOAC Int. 2005, 88, 1714. (30) Ben-Gigirey, B.; Rodriguez-Velasco, M. L.; Villar-Gonzalez, A.; Botana, L. M. J. Chromatogr., A In press. (31) Traynor, I. M.; Plumpton, L.; Fodey, T. L.; Higgins, C.; Elliott, C. T. J. AOAC Int. 2006, 89, 868-872. (32) Le Berre, M.; Kane, M. Anal. Lett. 2006, 39, 1587-1598. (33) Medina, M. B. J. Rapid Methods Autom. Microbiol. 2003, 11, 225-243. (34) Nedelkov, D.; Nelson, R. W. Appl. Environ. Microbiol. 2003, 69, 52125215. (35) Rasooly, A.; Herald, K. E. J. AOAC. Int. 2006, 89, 873-883. (36) Daly, S. J.; Keating, G. J.; Dillion, P. P.; Manning, B. M.; O’ Kennedy, R.; Lee, H. A.; Morgan, M. R. A. J. Agric. Food Chem. 2000, 48, 5097-5104. (37) Wiegele, J. B.; Barchi, R. L. Fed. Eur. Biochem. Soc. Lett. 1978, 91, 310314. (38) Powell, C. L.; Doucette, G. J. Nat. Toxins 1999, 7, 393-400. (39) Ruberu, S. R.; Liu, Y. G.; Wong, C. T.; Perera, S. K.; Langlois, G. W.; Doucette, G. J.; Powell, C. L. J. AOAC Int. 2003, 86, 737-745. (40) Llewellyn, L. E.; Doyle, J. Toxicon 2001, 39, 217-224. (41) Johnston, H. M.; Frey, P. A.; Angelotti, R.; Campbell, J. E.; Lewis, K. H. Proc. Soc. Exp. Biol. Med. 1964, 117, 425-430.

5908

Analytical Chemistry, Vol. 79, No. 15, August 1, 2007

for STX,43-45 neosaxitoxin,46,47 and gonyautoxin 2/3.21 Direct EIAs for STX48 and gonyautoxin 2/321 have also been described, as has a lateral flow immunochromatography (LFI) test.49 Although several EIA and LFI test kits are commercially available, their shortcomings can include detection of a limited number of PSP toxins at the required concentration, high cost, and potentially subjective interpretation of results. Given their strong potential for success based on the reports outlined above, the present study focused on evaluating the performance of three binders in a newly developed biosensor assay for PSP toxins: SCR derived from rat brains; a monoclonal antibody (GT13-A) raised to gonyautoxin 2/3;21 and, a polyclonal antibody raised to STX. MATERIALS AND METHODS Materials. Bovine serum albumin (BSA), 2,2-(ethylenedioxy)bis(ethylamine) (Jeffamine), sucrose, protease inhibitor cocktail (P8340), ethylenediaminetetraacetic acid (EDTA), potassium chloride, phosphate-buffered saline (PBS), 4-(2-hydroxyethyl)-1piperazineethanesulfonic acid (HEPES), 2-amino-2-hydroxymethyl1,3-propanediol (TRIS), 2-morpholinoethanesulfonic acid (MES), 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC), N-hydroxysuccinimide (NHS), formaldehyde and Freund’s complete and incomplete adjuvant were purchased from SigmaAldrich (Dorset, UK). Male Holtzman rat brains (6 weeks old) were purchased from Harlan Sera-Lab Ltd. Ethanolamine, HBSEP buffer (pH 7.4, 0.01 M HEPES, 0.15 M NaCl, 3 mM EDTA, 0.005% polysorbate), and an amine coupling kit were provided by Biacore AB (Uppsala, Sweden). STX diacetate (STXdiAc; 2.74 mg/ mL), STX dihydrochloride (STXdiH); 24.2 µg/mL), decarbamoyl STX (dcSTX; 20.4 µg/mL), gonyautoxin 2/3 (GTX2/3; 62.1 µg/ mL), decarbamoyl gonyautoxin 2/3 (dcGTX2/3; 51.4 µg/mL), gonyautoxin 5 (GTX5; 24.7 µg/mL), gonyautoxin 1/4 (GTX1/4; 58.0 µg/mL), neoSTX (NEO; 25.2 µg/mL), decarbamoyl neosaxitoxin (dcNEO; 10.4 µg/mL), and C1/2 toxins (C1/2; 70.84 µg/ mL), as certified standard reference material, were obtained from the Institute for Marine Biosciences, National Research Council, Halifax, Canada (http://imb-ibm.nrc-cnrc.gc.ca/crmp/). Production of Biosensor Chip Surface. The carboxymethylated surface of a CM5 certified grade biosensor chip (Biacore AB) was equilibrated to room temperature and activated using an amine coupling kit. Briefly, EDC and NHS from the kit were mixed (1:1, v/v) and applied to the chip surface for 30 min. Excess (42) Carlson, R. E.; Lever, M. L.; Lee, B. W.; Guire, P. E. In Seafood Toxins; Ragelis, E. P., Ed.; American Chemical Society: Washington, DC. 1984; pp 181-192. (43) Chu, F. S.; Fan, T. S. L. J. AOAC Int. 1985, 68, 13-16. (44) Cembella, A.; Parent, Y.; Jones, D.; Lamoureux, G. In Toxic Marine Phytoplankton; Grane´li, E., Sundstrom, B., Edler, L., Anderson, D. M., Eds.; Elsevier: New York, 1990; pp 339-344. (45) Micheli, L.; Di Stefano, S.; Moscone, D.; Palleschi, G.; Marini, S.; Coletta, M.; Draisci, R.; Quadri, F. D. Anal. Bioanal. Chem. 2002, 373, 678-684. (46) Chu, F. S.; Huang, X.; Hall, S. J. AOAC Int. 1992, 75, 341-345. (47) Bu ¨ rk, C.; Usleber, E.; Dietrich, R.; Ma¨rtlbauer, E. Food Agric. Immunol. 1995, 7, 315-322. (48) Usleber, E.; Schneider, E.; Terplan, G. Lett. Appl. Microbiol. 1991, 13, 275277. (49) Laycock, M. V.; Jellett, J. F.; Belland, E. R.; Bishop, P. C.; The´riault, B. L.; Russell-Tattrie, A. L.; Quilliam, M. A.; Cembella, A. D.; Richards, R. C. In Harmful Algal Blooms 2000; Hallegraeff, G. M.. Blackburn, S. I.. Bolch, C. J.. Lewis, R. J., Eds.; Intergovernmental Oceanographic Commission of UNESCO: Paris. 2001; pp 254-256.

solution was removed, and Jeffamine (20%) in borate buffer was added to the chip surface and allowed to react for 1 h, followed by immobilization of STXdiAc onto the amine surface via aminoamino coupling. The chip surface was deactivated by exposure to ethanolamine (1 M) for 30 min. The sensor chip surface was then washed with deionized water, dried using a stream of nitrogen gas, and stored desiccated at 4 °C when not in use. Selection and Production of Binders. Preparation of SCR Extract. Rat brain synaptosome membrane fractions containing SCRs were prepared using a modification of the procedure described by Hartshorne and Catterall.50 All procedures were carried out on ice. Briefly, rat brains (with medulla removed) were each held in ice-cold buffer (pH 7.4; 12.5 mL) containing 0.32 M sucrose, 5 mM TRIS-HCl and protease inhibitor cocktail. Each brain was homogenized in the 0.32 M sucrose/TRIS buffer in a glass Teflon homogenizer (Wheaton, Fisher Scientific) controlled with an overhead stirrer (model SS10, Stuart, Fisher Scientific) at 75% full speed. The homogenized brains were suspended in 0.32 M sucrose/TRIS buffer (final volume, 200 mL) and rehomogenized. The suspension was centrifuged using a Sigma 3K30C centrifuge (700g, 10 min, 4 °C). The supernatant was retained and the pellet resuspended in 0.32 M sucrose/TRIS buffer (134 mL) and centrifuged (700g, 10 min, 4 °C). The pellet was discarded, and both supernatants were combined and centrifuged (27000g, 40 min, 4 °C). The supernatant was discarded and the pellet osmotically lysed by suspension in 5 mM TRIS-HCl (pH 8.2) (333 mL), containing 1 mM EDTA and protease inhibitor cocktail, incubated on ice for 15 min, and homogenized with four strokes of the glass Teflon homogenizer. The lysed membrane homogenate was centrifuged (27000g, 40 min, 4 °C) and the pellet resuspended in a buffer (pH 7.4) containing 200 mM potassium chloride, 10 mM HEPES/TRIS, and the protease inhibitor cocktail (final volume, 40 mL). The SCR preparation was aliquoted (500 µL) and stored at -80 °C. Monoclonal Antibody (GT13-A). Monoclonal antibody GT13-A raised to gonyautoxin 2/3 (GTX2/3)-keyhole limpet hemeocyanin (KLH) protein conjugate was kindly supplied as a gift (Dr. K. Kawatsu, Division of Bacteriology, Osaka Prefectural Institute of Public Health, Osaka, Japan). The synthesis of the GTX2/3-KLH immunogen, the immunization process, antibody titer determination, assessment of antibody sensitivity, and specificity for GT13-A by enzyme immunoassay have been described in detail elsewhere.21 Production of a Polyclonal Antibody (R895). Preparation of Immunogen. A STX-Jeffamine-BSA immunogen was synthesized. STX was conjugated to the carrier protein, BSA via the Mannich reaction incorporating Jeffamine as a spacer compound. In brief, Jeffamine was conjugated initially to the BSA via amine coupling. BSA (10 mg) was dissolved in MES buffer (0.5 mL, 0.05 M, 0.5 M NaCl, pH 5). An aliquot (250 µL) of EDC (20 mg) and NHS (8 mg) dissolved in MES buffer (2 mL, 0.05 M, 0.5 M NaCl, pH 5) was added to the BSA and mixed for 5 min at room temperature. Then Jeffamine (50 µL, 1 M) was added and the mixture allowed to react for 3 h at room temperature. The Jeffamine-BSA conjugate (800 µL) was purified using a PD-10 column (GE Healthcare) versus PBS (pH 7.2). The Jeffamine-BSA (10 mg of protein) was eluted using PBS (3.5 mL, pH 7.2). This eluant was (50) Hartshorne, R. P.; Catterall, W. A. J. Biol. Chem. 1984, 259, 1167-1675.

split into 10 × 350 µL (10 × 1 mg) aliquots and freeze-dried. Freeze-dried Jeffamine-BSA (1 mg) was resuspended in 350 µL of water to give 1 mg of protein in PBS pH 7.2. STX (200 µg; 73 µL of 6524 µM) and 2.5% formaldehyde (in water) (80 µL) were then added. This mixture was allowed to react for 50 h followed by dialysis over 24 h in 3 × 4 L of 0.15 M saline solution. Immunization of Rabbits. The STX-Jeffamine-BSA immunogen in sterile saline (180 µg/180 µL) was emulsified with Freund’s complete adjuvant containing heat-killed Mycobacterium tuberculosis and injected subcutaneously into a New Zealand White female rabbit. Subsequent booster injections of immunogen (60 µg/60 µL) using Freund’s incomplete adjuvant were administered on a monthly basis. Test bleeds were collected 10 days after each booster injection and monitored for both the presence and specificity of antibodies for the detection of PSP toxins with the STX chip surface using a Biacore Q biosensor with control and evaluation software (Biacore AB). Briefly, dilutions of serum in HBS-EP buffer were injected over the chip surface to establish whether antibody binding to the STX on the surface was occurring. If binding to the surface was obtained, diluted serum and STX standard were injected together over the chip surface to establish the specificity of this binding to the surface compared to the STX standard in solution. Following confirmation of significant antibody titer from the test bleed subsequent to the third booster injection, demonstrated by both binding to the STX chip surface and inhibition of binding in the presence of PSP toxins, harvesting of the antibody was performed 10 days following the fourth and final booster injection. After harvesting, a full evaluation of the sensitivity and specificity of the antibody for PSP toxins was determined by biosensor analysis. Determination of the Binding Capacity of Each Binder to the STX Chip Surface. The maximum binding capacity (Rmax) of each of the three binders (i.e., SCR, GT13-A, R895) to STX immobilized on the chip surface was measured using the Biacore Q biosensor. For the SCR, this was determined by a 20-min injection (5 µL/min) of a 1/4 dilution of the SCR extract in HBSEP buffer. For each of the antibodies, a 1/10 dilution of binder was applied in the HBS-EP buffer at the same flow rate. Assessment of Sensitivity and Specificity of Each Binder to the STX Surface Using the Biacore Q Assay. The sensitivity and specificity of each binder (i.e., SCR, GT13-A, R895) to STX and its analogues were assessed using the Biacore Q biosensor with the STX surface. Stock solutions of STXdiAc, STXdiH, dcSTX, GTX2/3, dcGTX2/3, GTX5, GTX1/4, NEO, dcNEO, and C1/2 were used to prepare working standards in pH 7.4 HBS-EP buffer. A set of working standards for each toxin, ranging in concentration from 0 to 200 ng/mL, were prepared to produce calibration curves for each toxin based on dose-response on the biosensor. Various parameters were investigated initially in developing the biosensor assay including: binder dilution, ratio of binder to standard, flow rate, injection volume, and contact time. The optimum parameters determined for each binder and used subsequently to assess their sensitivity and specificity to the STX chip surface are summarized in Table 1. In brief, each PSP working standard was mixed with a dilution of binder in HBS-EP buffer and injected over the sensor chip surface at a flow rate of 25 µL/min. Report points were recorded before (5 s) and after each injection (30 s), and the relative Analytical Chemistry, Vol. 79, No. 15, August 1, 2007

5909

response units were determined. The chip surface was regenerated with a 25-µL injection of sodium hydroxide (100 mM) at a flow rate of 25 µL/min. However, to regenerate the chip surface when using the SCR as a binder, acetonitrile (20%) in NaOH solution (250 mM) (25 µL) was injected over the chip surface at a flow rate of 25 µL/min. A typical analytical cycle was completed in ∼8 min and each standard was analyzed in duplicate. Calculation of Binder Sensitivity. The average relative response unit of the 0 ng/mL standard (HBS-EP buffer pH 7.4, only) represented 100% binding of a binder to the chip surface, and the average response units for the remaining standard concentrations were normalized relative to the response units of the 0 ng/mL standard. The percentage binding with each of the standards was then calculated as follows:

% binding ) (response units of standard/ response units of 0 ng/mL standard) × 100 The midpoint of each standard curve was determined from the following equation in order to establish the IC50 for each PSP toxin for each binder:

midpoint of curve ) ((100% binding (i.e., % binding of 0 ng/mL standard) % binding of highest standard concentration)/2) + (% binding of highest standard concentration) IC50 is defined as the concentration of a PSP toxin required to reduce the response to 50% binding compared to the response when no toxin is present (100% binding). Calculation of Binder Specificity. The cross-reactivity profile of each binder to the PSP toxin standards was calculated from the IC50 in the biosensor assay as follows:

cross-reactivity ) (IC50 of STX/IC50 of PSP toxin) × 100 Calculation of Dynamic Range (IC10 - IC90). IC10 and IC90 are defined as the concentrations of a PSP toxin required to reduce the response to 10 and 90% binding, respectively, compared to the response when no toxin is present (100% binding). The concentration range determined from these two points is considered to be the dynamic range of the assay and was calculated as follows:

IC90 point on curve ) ((100% binding % binding of highest standard concentration)/0.9) + (% binding of highest standard concentration) Whereas,

IC10 point on curve ) ((100% binding % binding of highest standard concentration)/0.1) + (% binding of highest standard concentration) RESULTS AND DISCUSSION Rmax of Binders to the STX Chip Surface. The maximum binding capacities (Rmax) of the SCR, GT13-A antibody, and R895 5910 Analytical Chemistry, Vol. 79, No. 15, August 1, 2007

Table 1. Optimized Biacore Q Biosensor Parameters for Each Binder binder parameter

GT13-A

SCR

R895

binder dilution in HBS ratio of binder to PSP standard flow rate, µL/min injection volume, µL contact time, min

1/250 1:4 v/v 25 50 2

1/2 1:1 v/v 25 25 1

1/250 1:4 v/v 25 50 2

antibody to the STX immobilized on the chip surface were determined to be 2500, >6000, and >8000 response units, respectively. This result established that molecular interactions were indeed occurring between the binders and the STX immobilized on the chip surface. For efficient and cost-effective analysis using optical biosensing methods, an essential requirement is regeneration of the sensor chip surface for removal of all noncovalently bound material while maintaining the structural integrity of the attached ligand. For both monoclonal and polyclonal antibodies, NaOH (100 mM) achieved total removal of proteins without causing deterioration of the immobilized STX. In the case of SCRs, acetonitrile (20% aqueous), in addition to a higher molarity of NaOH (250 mM), were required to effectively remove this binder from the surface. Determination of the Sensitivity of Each Binder to the STX Surface Using a Biacore Q Biosensor Assay. To determine the sensitivity of the SCR, and GT13-A and R895 antibodies against STX, standard calibration curves of STXdiAc concentration versus response for each binder were generated. Figure 2 shows the normalized standard calibration curves produced for each binder in the presence of STXdiAc standards ranging from 0 to 1000 ng/ mL in HBS-EP buffer (pH 7.4). The maximum percent inhibition of binding of the SCR, GT13A, and R895 to the STX chip surface over this standard range was 27.8, 93.3, and 87.2%, respectively. The corresponding IC50 values for each binder determined from these standard calibration curves were as follows: SCR 8.11 ng/mL, GT13-A 5.77 ng/mL, and R895 1.56 ng/mL. The sensitivity of the SCR-based assay was thus lower than both antibody-based assays as well as the radioreceptor assay published previously, which also employed a sodium channel preparation.8 In contrast, the GT13-A EIA exhibited an IC50 of 4.06 ng/mL for STXdiAc,21 a sensitivity very similar to the biosensor-based protocol using the same monoclonal antibody. Nevertheless, the polyclonal binder (R895) clearly demonstrated the highest sensitivity to saxitoxin in the present biosensor format. Determination of the Specificity of Each Binder against STX Analogues. The specificity of each binder was assessed by comparing the maximum binding efficiency of the binder to the STX chip surface in the presence of all available PSP toxin standards over a range of concentrations. Figure 3 illustrates the cross-reactivity binding profile of the SCR extract to STXdiAc, STXdiH, GTX2/3, dcGTX2/3, GTX5, NEO, and GTX1/4. The maximum percent inhibition of binding to the chip surface was relatively low compared to that of the antibodies. NEO displayed the greatest maximum inhibition of binding (36.1%) for the SCR. Moreover, calibration curves were shallow, which is often associated with problems relating to assay

Figure 2. Comparison of the SCR preparation, GT13-A monoclonal antibody, and R895 polyclonal antibody normalized calibration curves obtained for STXdiAc certified reference standard concentrations from 0 to 1000 ng/mL in HBS-EP buffer (pH 7.4).

Figure 3. SCR preparation cross-reactivity profile from normalized calibration curves obtained using available PSP toxin standards at concentrations from 0 to 100 ng/mL in HBS-EP buffer (pH 7.4).

accuracy and reproducibility when used for routine testing. Each toxin tested produced a similar response with the SCR binder. Consequently, the dynamic ranges of the calibration curves were lower than required to accurately quantify the relative crossreactivities. Further technical difficulties were encountered with the SCR. The SCR preparation appeared to be stable for only a short time at 4 °C, with the molecular interactions between the SCR and chip

surface declining by >50% within several hours. To overcome this problem, freshly thawed aliquots of frozen material were diluted in buffer and stored briefly at 4 °C until injection. In addition, the need to modify/strengthen the regeneration solution to completely remove bound SCR from the chip surface was also considered to be a disadvantage, with performance of the chip surface deteriorating after only 20-30 injections. The basic NaOH solution (250 mM) used to regenerate the chip surface likely altered the surface Analytical Chemistry, Vol. 79, No. 15, August 1, 2007

5911

Figure 4. GT13-A antibody cross-reactivity profile from normalized calibration curves obtained using available PSP standards at concentrations from 0 to 200 ng/mL in HBS-EP buffer (pH 7.4).

Figure 5. R895 antibody cross-reactivity profile from normalized calibration curves obtained using available PSP standards at concentrations from 0 to 200 ng/mL in HBS-EP buffer (pH 7.4).

charge of the immobilized saxitoxin, causing a decline in assay performance. Figures 4 and 5 show the cross-reactivity profiles of the binding of the GT13-A and R895 antibodies to STXdiAc, STXdiH, dcSTX, GTX2/3, dcGTX2/3, GTX5, C1/2, dcNEO, NEO, and GTX1/4. Table 2 presents the IC50 values for each toxin and calculated percent cross-reactivity values in relation to STX. In addition, the corresponding data extracted from the EIA publication that produced the GT13-A EIA are also shown. 5912

Analytical Chemistry, Vol. 79, No. 15, August 1, 2007

Results obtained using the GT13-A monoclonal antibody as a binder showed a very clear pattern. While excellent IC50 values were obtained with the non-N1-hydroxylated PSP toxins (i.e., H atom at the R1 position), this antibody did not bind the N1hydroxylated toxins dcNEO, NEO, or GTX1/4 (i.e., OH group at the R1 position) to a significant extent. The cross-reactivity values at the IC50 point obtained previously (EIA)21 for GT13-A against GTX2/3, dcGTX2/3, NEO, and GTX1/4 were compared to those using the same antibody in the biosensor assay (Table 2).

Table 2. GT13-A and R895 Biosensor Assay Cross-Reactivity Data to STXdiAc Compared with Published GT13-A EIA Data21 biosensor assay with STX Chip surface GT13-Aa PSP toxin STXdiAc STXdiH dcSTX NEO dcNEO GTX 1/4 GTX 2/3 C1/C2 dcGTX 2/3 GTX 5 a

Table 3. GT13-A and R895 Biosensor Assay Dynamic Ranges for All Saxitoxin Analogues Tested dynamic range IC90 to IC10 (ng/mL)

EIA

R895a

IC50

% crossreactivity

IC50

% crossreactivity

5.68 4.73 3.61 94.74 86.30 80.88 3.59 3.36 2.72 4.27

100 124.1 162.7 6.2 6.8 7.3 163.6 174.9 216.1 137.6

1.56 1.47 1.55 7.81 36.85 65.91 3.58 9.79 16.10 2.53

100 105.9 100.7 19.9 4.2 2.4 43.4 15.9 9.7 61.4

GT13-A IC50

% crossreactivity

4.06 4.06

100 100

89.37

4.5

3.46 0.28

117 1450

0.41

990

Toxin standard range, 0-200 ng/mL.

Although cross-reactivity values for the EIA assay were higher, the same trend was obtained in both assays with NEO showing the weakest cross-reaction. The low cross-reactivity values for NEO and GTX1/4 in the biosensor assay (e7.3) may be associated with the hydroxyl group at the R1 position of these molecules (R1 ) OH, Figure 1), which could modify their hydrophobicity or charge, or cause steric hindrance in the binding of GT13-A to NEO and GTX1/4 compared to the non-N1-hydoxylated forms. Use of the R895 polyclonal antibody in the biosensor assay yielded generally lower IC50 values and a narrower range of percent cross-reactivity values compared to those obtained with GT13-A as the binder. These data showed that the R895 biosensor assay bound all PSP toxins tested, yet cross-reactivity with dcNEO and GTX1/4 remained problematic. However, unlike the GT13-A biosensor assay, NEO, the more potent toxin, was detectable at low concentrations. The high IC50 values obtained for dcNEO and GTX1/4 with the R895 biosensor may reflect the combination of either a hydroxyl group at both R1 and R4 positions of the toxin molecule (R1 ) R4 ) OH, Figure 1) or the combination of an OH at the R1 position along with sulfate groups at positions R2 or R3, respectively. As with the GT13-A antibody, these substitutions may change the chemical or steric properties of the toxin molecules enough to affect the binding affinity of R895. Less than optimal cross-reactivities with the N1-hydroxylated PSP toxins have been reported frequently for antibody-based detection methods.48,49 EC Directive 91/492/EEC indicates that the total PSP toxin content must not exceed 80 µg STX equivs/100 g of shellfish flesh. Based on the different extraction protocols for the AOAC mouse bioassay and the AOAC Lawrence HPLC-FLD method (prior to oxidation), this action level equates to toxin concentrations of 400 and 100 ng/mL, respectively. The cross-reactivity profiles over the dynamic ranges (IC90 to IC10) of both the GT13-A and R895 antibody standard buffer curves (Table 3) indicate that the majority of PSP toxins are detectable at these concentrations. Use of the AOAC mouse bioassay extraction procedure would permit detection of all PSP toxins within the critical concentrations by the biosensor assay employing either antibody. However, this method produces a relatively crude extract and matrix effects may affect assay performance to the extent that further sample cleanup may

PSP toxin

GT13-A

R895

STXdiAc STXdiH dcSTX NEO dcNEO GTX1/4 GTX2/3 C1/C2 dcGTX2/3 GTX 5

1.5-39.6 0.7-36.8 0.5-14.5 15.2-179.0 13.6-176.4 14.2-174.6 0.6-9.0 0.5-8.1 0.4-6.3 0.8-16.9

0.4-3.6 0.3-3.0 0.3-4.8 0.7-87.6 2.7-135.3 2.9-172.7 0.5-44.0 0.6-90.4 1.1-96.3 0.5-8.4

be necessary. For samples prepared according to the Lawrence technique, which includes a C18 solid-phase extraction clean-up step, the GT13-A antibody may not detect dcNEO, NEO, and GTX1/4 at or below the critical concentrations. By comparison and based on its dynamic ranges (Table 3), the R895 antibody showed the potential to detect all important PSP toxins within the critical range, but extensive validation studies will be required and detecting dcNEO and GTX1/4 may still prove difficult. An alternative approach to enhancing the performance of the biosensor assays would involve the development of a dedicated extraction protocol. CONCLUSIONS The present study has demonstrated that a stable optical biosensor chip surface can be produced using immobilized STX. This surface was capable of molecular interactions with a range of PSP toxin binding proteins, each of which exhibited different sensitivity and cross-reactivity profiles. While the SCR binder displayed the most promising cross-reactivity profile, a number of important technical difficulties relating to reproducibility and reliability were encountered in using this rat brain preparation. Although considerable effort was applied to overcome these complications (data not reported), the authors concluded that further development of a robust biosensor assay using this binder would require development of an improved binder extraction procedure, alternative surface chemistry, and/or a more gentle regeneration solution. The two antibodies characterized in this study yielded encouraging results. Both binders achieved the reproducibly high sensitivity required for development of a robust, routine assay; nonetheless, a number of PSP toxins may not be detected within the concentration range required for regulatory monitoring. This was true particularly of GT13-A, which was unable to detect NEO, one of the most potent PSP toxins, within the range of concentrations required. Research is under way to develop additional binders and to modify the assay formats described herein, with the aim of delivering an optical biosensor assay capable of replacing the mouse bioassay. Safety. Saxitoxin and its analogues are responsible for incidents of paralytic shellfish poisoning. Therefore, when using PSP toxin standard solutions, special care should be taken. Gloves and Analytical Chemistry, Vol. 79, No. 15, August 1, 2007

5913

eye protection should be worn at all times. Appropriate disposal methods should also be utilized. ACKNOWLEDGMENT This work was supported by an EU research grant BioCop Contract FOOD-CT-2004-06988. NOS does not approve, recommend, or endorse any proprietary product or material mentioned in this publication. No reference shall be made to NOS, or to this publication furnished by NOS, in any advertising or sales promotion which would indicate or imply that NOS approves, recom-

5914

Analytical Chemistry, Vol. 79, No. 15, August 1, 2007

mends, or endorses any proprietary product or proprietary material mentioned herein or which has as its purpose any intent to cause directly or indirectly the advertised product to be used or purchased because of NOS publication.

Received for review February 18, 2007. Accepted April 11, 2007. AC070342O