Specificity of Bacillus thuringiensis for lepidopteran larvae: factors involved in vivo and in the structure of a purified protoxin.

The relative LD50 values in two test Lepidoptera of Bacillus thuringiensis subspecies kurstaki HD1, which contains three crylA protoxin genes, was the same as a plasmid-cured derivative or a Bacillus cereus transcipient containing only one of the three genes. Differential rates of transcription of these genes in the original strain could account, at least partly, for this result. Strains containing only the single protoxin gene (crylA(b] produced inclusions when grown at 25 degrees C but not 32 degrees C, despite transcription of this gene at both temperatures. The instability of the crylA(b) protoxin was not found in the parental B. thuringiensis subsp. kurstaki HD1 strain grown at either temperature, however, so kurstaki HD1 strains with multiple protoxin genes must produce some stabilizing factor, perhaps another protoxin. The cryl protoxins contain a highly conserved carboxyl half which is proteolytically removed upon conversion to toxin. All of the protoxin cysteines are present in protease-sensitive regions and they are oxidized in inclusions. Most of the disulphides appear to be essential for specificity since their reduction in the crylA(b) protoxin resulted in loss of selectivity for one of the test insects. This lack of specificity was also found for this protoxin produced by an Escherichia coli clone, probably because of the reducing conditions in these cells. Specificity was restored by reoxidation of the pure protoxin, by removal of the carboxyl half of oxidized protoxin with trypsin, or by subcloning of the toxin portion. The oxidized form of protoxins must be important for specificity, for the formation of crystalline inclusions, and probably for interactions required for the stabilization of some protoxins.