two fractions, purified through the third TLC, appeared to contain PbTx-2 by HPLC 

 separation and co-elution. Radioimmunoassay was performed on these five fractions, using 

 tritiated PbTx-3 as the internal displacement standard. Based on this assay, the three 

 fractions purified through 2 TLC steps contained 10.2, 12.2, and 9.33 ng toxin/g liver; the two 

 fractions purified through 3 TLC steps contained 17 and 240ng toxin/g liver. 



The process of extraction, purification, chromatographic separation, and 

 radioimmunoassay conducted on the 50 samples led to the conclusion that five of the 

 samples contained brevetoxin or some very similar toxin. Reported concentrations in original 

 samples were calculated by proportion of sub-sampling at the various steps and were based 

 on "PbTx-3 equivalents" in the radioimmunoassay. 



Of the five toxin-spiked control samples only one was detected as containing 

 brevetoxin; this sample was spiked with the largest amoimt of PbTx-3, 25ug. Two other 

 samples were spiked with 20 and 15ug of PbTx-3 respectively, but were not identified as 

 containing brevetoxin. PbTx-1 and PbTx-2 were also added to several of the samples; PbTx-1 

 is known to hydrolyze quite quickly. The fact that purified toxins "stick" to glass- and plastic- 

 ware may expleiin the low level of apparent spike of the liver samples. It is quite possible 

 that neither the PbTx-1 or PbTx-2 spikes were effective, or it is possible that they do not 

 effectively displace radio-labeled PbTx-3 in the radioiiim[iunoassay. 



Of the five carrier-spiked control samples (treated with MeOH only), three were 

 identified by the radioimmunoassay as containing brevetoxin. It is difficult to explain this 

 finding. The other two carrier-spiked samples were found to be negative when purified to 

 the second TLC step. It is possible that an interfering substance was removed in the early 

 cleanup phases of some of the controls and not in others. 



The sample reported to contain the largest cmiount of brevetoxin, as determined by 

 radioimmunoassay, was one of the non-toxin (MeOH only) spiked control samples. The only 

 dolphin liver sample from the strandings that was identified as containing brevetoxin at all 

 stages contained 10.2 ng toxin/g liver. This level of toxin is considered to be very low. 



The problems encountered in properly identifying the spiked and non-spiked control 

 samples raised serious questions concerning the efficacy of this assay method for detecting 

 brevetoxin in bottlenose dolphin liver samples. The assay of the control samples resulted in 

 both false-positives and false-negatives. Certainly, the results of this brevetoxin analysis 

 caimot be considered conclusive. That is, based on the incorrect assay results of the control 

 samples, brevetoxin poisoning cannot be ruled out as a proximate cause or factor in the 1990 

 bottlenose dolphin strandings. These results also indicate that other studies (e.g., Geraci 

 1989) of brevetoxin poisoning in bottlenose dolphins which have employed similar assay 

 methods without adequate controls (both known positives and known negatives) should also 

 be considered inconclusive. 



54 



