You are here:
TOXICITY ASSESSMENT OF PARALYTIC SHELLFISH POISONS (PSPS) USING QUANTITATIVE STRUCTURE-ACTIVITY RELATIONSHIPS
DELACRUZ, A. A., B. BOUTIN, AND R. VENKATAPATHY. TOXICITY ASSESSMENT OF PARALYTIC SHELLFISH POISONS (PSPS) USING QUANTITATIVE STRUCTURE-ACTIVITY RELATIONSHIPS. Presented at AOAC Annual Meeting, Orlando, FL, September 11 - 15, 2005.
The major objective for this task is to develop analytical methods to detect problematic cyanotoxins in water. A preconcentration/extraction procedure will be initially developed to remove interfering substances with the detection of cyanotoxins in various water matrices. These methods will be used to determine the occurrence/prevalence of cyanotoxins in water and in the cyanobacterial cell culture collection in-house, and compare with the effectiveness of innovative detection methods developed for biological toxins.
The most significant harmful algal bloom (HAB) toxin in terms of public health is commonly known as paralytic shellfish poisons (PSPs, "red tides" toxins). PSPs are neurotoxins produced by marine dinoflagellates and some cyanobacteria. PSPs comprise of over 21 natural toxins with varying toxicity and are broadly classified into saxitoxins, neosaxitoxins, gonyautoxins, decarbamoyl gonyautoxins and decarbamoyl saxitoxins, depending upon the functional groups R1, R2, R3 and R4 (see Figure below). Humans potentially may be exposed to PSP poisoning through drinking water, recreational water and hemodialysis treatment. Consumption of fish and shellfish has been implicated in human outbreaks of PSP in the US and worldwide. Humans have been reported to exhibit characteristic neurological symptoms, nausea, vomiting, diarrhea and sometimes death within 2-12 hours. Standard mouse bioassay data are available in the literature for some PSPs. The mouse toxicity model is labor-intensive, time-consuming and expensive. Routine use of animal models to detect and quantify PSP toxicity may be circumvented using Quantitative Structure-Activity Relationships (QSARs). Currently, there are no QSARs in the literature for predicting the toxicity of PSPs. USEPA is investigating the use of QSAR models to support risk assessment of chemical toxicity. The objective of the study is to develop a QSAR model that can provide quantitative estimates of the PSP dose that causes toxicity in mice. To meet the objective, known PSP mouse bioassay data were collected from the literature. The molecular structures of the PSPs were optimized to their global minimum using the CONFLEX module in CAChe (Computer Aided Chemistry) software. The descriptor generator programs Dragon, Molconn-Z, CAChe and AMPAC/CODESSA were used to generate the descriptors for the QSAR model. Initially, a QSAR model for PSP toxicity was developed using 2-dimensional descriptors from Dragon as independent variables. The QSAR models had a high degree of predictability using the Dragon descriptors. However, the models predicted similar toxicities for some of the isomeric PSPs and conformers as the 2-dimensional descriptors could not differentiate between their respective chemical structures. The models were redeveloped with 3-dimensional descriptors from CAChe and AMPAC/CODESSA. The results of the QSAR model using 2-dimensional and 3-dimensional descriptors will be presented and discussed.