Pore forming toxins rely on oligomerization for membrane insertion to kill their targets. Bacillus thuringiensis produces insecticidal Cry-proteins composed of three domains that form pores that kill the insect larvae. Domain I is involved in oligomerization and membrane insertion, whereas Domains II and III participate in receptor binding and specificity. However, the structural changes involved in membrane insertion of these proteins remain unsolved. The most widely accepted model for membrane insertion, the 'umbrella model', proposed that the α-4/α-5 hairpin of Domain I swings away and is inserted into the membrane. To determine the topology of Cry1Ab in the membrane, disulfide bonds linking α-helices of Domain I were introduced to restrict their movement. Disulfide bonds between helices α-2/α-3 or α-3/α-4 lost oligomerization and toxicity, indicating that movement of these helices is needed for insecticidal activity. By contrast, disulfide bonds linking helices α-5/α-6 did not affect toxicity, which contradicts the 'umbrella model'. Additionally, Föster resonance energy transfer closest approach analyses measuring distances of different points in the toxin to the membrane plane and collisional quenching assays analysing the protection of specific fluorescent-labeled residues to the soluble potassium iodide quencher in the membrane inserted state were performed. Overall, the data show that Domain I from Cry1Ab may undergo a major conformational change during its membrane insertion, where the N-terminal region (helices α-1 to α-4) participates in oligomerization and toxicity, probably forming an extended helix. These data break a paradigm, showing a new 'folding white-cane model', which better explains the structural changes of Cry toxins during insertion into the membrane.
Read full abstract