A collaborative study was conducted on a microplate format receptor binding assay (RBA) for paralytic e shellfish toxins (PST). The assay quantifies the composite PST toxicity in shellfish samples based on the ability of sample extracts to compete with (3)H saxitoxin (STX) diHCl for binding to voltage-gated sodium channels in a rat brain membrane preparation. Quantification of binding can be carried out using either a microplate or traditional scintillation counter; both end points were included in this study. Nine laboratories from six countries completed the study. One laboratory analyzed the samples using the precolumn oxidation HPLC method (AOAC Method 2005.06) to determine the STX congener composition. Three laboratories performed the mouse bioassay (AOAC Method 959.08). The study focused on the ability of the assay to measure the PST toxicity of samples below, near, or slightly above the regulatory limit of 800 (microg STX diHCl equiv./kg). A total of 21 shellfish homogenates were extracted in 0.1 M HCl, and the extracts were analyzed by RBA in three assays on separate days. Samples included naturally contaminated shellfish samples of different species collected from several geographic regions, which contained varying STX congener profiles due to their exposure to different PST-producing dinoflagellate species or differences in toxin metabolism: blue mussel (Mytilus edulis) from the U.S. east and west coasts, California mussel (Mytilus californianus) from the U.S. west coast, chorito mussel (Mytilus chiliensis) from Chile, green mussel (Perna canaliculus) from New Zealand, Atlantic surf clam (Spisula solidissima) from the U.S. east coast, butter clam (Saxidomus gigantea) from the west coast of the United States, almeja clam (Venus antiqua) from Chile, and Atlantic sea scallop (Plactopecten magellanicus) from the U.S. east coast. All samples were provided as whole animal homogenates, except Atlantic sea scallop and green mussel, from which only the hepatopancreas was homogenized. Among the naturally contaminated samples, five were blind duplicates used for calculation of RSDr. The interlaboratory RSDR of the assay for 21 samples tested in nine laboratories was 33.1%, yielding a HorRat value of 2.0. Removal of results for one laboratory that reported systematically low values resulted in an average RSDR of 28.7% and average HorRat value of 1.8. Intralaboratory RSDr based on five blind duplicate samples tested in separate assays, was 25.1%. RSDr obtained by individual laboratories ranged from 11.8 to 34.9%. Laboratories that are routine users of the assay performed better than nonroutine users, with an average RSDr of 17.1%. Recovery of STX from spiked shellfish homogenates was 88.1-93.3%. Correlation with the mouse bioassay yielded a slope of 1.64 and correlation coefficient (r(2)) of 0.84, while correlation with the precolumn oxidation HPLC method yielded a slope of 1.20 and an r(2) of 0.92. When samples were sorted according to increasing toxin concentration (microg STX diHCl equiv./kg) as assessed by the mouse bioassay, the RBA returned no false negatives relative to the 800 microg STX diHCl equiv./kg regulatory limit for shellfish. Currently, no validated methods other than the mouse bioassay directly measure a composite toxic potency for PST in shellfish. The results of this interlaboratory study demonstrate that the RBA is suitable for the routine determination of PST in shellfish in appropriately equipped laboratories.