RADIOIMhfUNQASSAY 



This procedure wis not within the oontncted SOW, and wis bsted in the Fee Schedule as ar 

 Additional Ana)ytical Protocol. However, based on the hmhed number of samples reaching HPLC 

 immunoassay of leveral samples wis undertaken to confirm identity, at our expense. At present, only 

 brevetoxins and dguatoxin are known to cross-react in this radioimmunoassay. The %amp)es tested were: 



CC094 (podtivt through two TLC pises, negative by HPLQ 



2505-5 (positive through two TLC plates— two fractions, negative by HPLQ 



2505-6 (positive through two TLC ptates, negative ly HPLQ 



2505-9 (positive through three TLC plates— four fractions, positive Yry HPLQ 



2505-10 (positive through three TLC plates, positive by HPLQ 



T^ resuhs of RL\ are flhistrated in Rgures 1-5. An intemal standard displacement carve for 

 unlabeled PbTx-3 as compcihor (against fixed trttiated PbTx-3) is used in each experiment. Based on these 

 resuhs, and we feel these are unequivocal, each of the five samples indicated above was oontammated with 

 brevetoxin to one degree or another. Based on RIA, the amooirt of brevttoxin contammatmg the ongmaJ 

 samples was calculated to be as follows: 



cam 0390 Mg per 38i)36 g, or 102 Bg^ liver 



2505-5 0.460 Mg per 41.255 g, or 12.1 ng^ bver 



2505-6 0385 Atg per 43.482 g, or 933 ag^ liver 



2505-9 im Mg per 46.639 g. or 240 Bg^ liver 



2505-10 0.815 Mg per 45332 g, or 17jO ng/g bver 



POTENTIAL SOURCES OF ERROR 



We antic^>ated and were informed that random brevetoxin-spiked samples were induded amongst the 

 50 samples submitted for analysis, and that^amples fiom the previous Atlantic do^hin die-off of 1987-1988 

 might also be induded. In the former instance, that of ^iked samples, we expea to have little dif&uhy in 

 identifying PbTx-2 and PbTx-3 (fairiy stable materials) and greater difBcuhy with Pbli-l or aiy of its oogeners 

 (an unstable polyether backbone structure). We in fact expntsed our uncertainty with respeo to toxins with 

 the PbTx-1 backbone prior to contract issuance. With reqject to samples from the previous Atlantic bottlenose 

 die-off, we expressed a concern for the length of time in storage, whidi if it was not at -80*C might not be 

 sufficiently cold to inhibh aO enzymatic activity. Our previous experience with a progressive reduction in 

 dguatoxicity in shark bver is our basis for this concern. In these latter two cases, we might ei p c fl to tee less 

 or no brevetoxin in samples, whether q>iked or authentic 



Trace brevetoxin contamination within our R & D laboratories is a potential problem, there being 

 b r ev e toxin purification progr e ssin g at any time. To alleviate the potential for oontamiiuaion, pristine glassware 

 was purchased for these analyses, and was used only once. Thin-layer chromatography plates were newly 

 unopened boxes and soWents were freshly opened prior to use. HPLC syringes were sequentially rinsed in an 

 duotrophic series of soWents as were aniJytical HPLC cohimns prior to use. Fixed baseline at the absorbance 

 sensitivity used for analyses were a pre-requisite. Ehition times which match authentic brevetoxin retention 

 times are circumstantial and not ifieatiBactioa means in an authentic sense. Nor is oo-migration of authentic 

 and unknown in the same sample. However, co-migration lends an iitcreased measure of certainty to the 

 conclusions. Sensitivity of deteaion is prinqpaOy a function of ultraviolet ab so rb an ce, and this decreases from 

 PbTx-2 to PbTx-3 to PbTx-9, and from PbTx-1 to PbTx-7 to PbTx-10. Thus, h wiD be more difficuh to detect 

 PbTx-9 and PbTx-10, less difficuh to see PbTx-7 and PbTx-3, and relatively easy to see PbTx-1 and PbTx-2. 



125 



