Computer modeling of binding of diverse weak toxins to nicotinic acetylcholine receptors

Abstract

Weak toxins are the “three-fingered” snake venoms toxins grouped together by having an additional disulfide in the N-terminal loop I. In general, weak toxins have low toxicity, and biological targets have been identified for some of them only, recently by detecting the effects on the nicotinic acetylcholine receptors (nAChR). Here the methods of docking and molecular dynamics simulations are used for comparative modeling of the complexes between four weak toxins of known spatial structure (WTX, candoxin, bucandin, γ-bungarotoxin) and nAChRs. WTX and candoxin are those toxins whose blocking of the neuronal α7- and muscle-type nAChR has been earlier shown in binding assays and electrophysiological experiments, while for the other two toxins no such activity has been reported. Only candoxin and WTX are found here to give stable solutions for the toxin-nAChR complexes. These toxins appear to approach the binding site similarly to short α-neurotoxins, but their final position resembles that of α-cobratoxin, a long α-neurotoxin, in the complex with the acetylcholine-binding protein. The final spatial structures of candoxin and WTX complexes with the α7 neuronal or muscle-type nAChR are very similar and do not provide immediate answer why candoxin has a much higher affinity than WTX, but both of them share a virtually irreversible mode of binding to one or both these nAChR subtypes. Possible explanation comes from docking and MD simulations which predict fast kinetics of candoxin association with nAChR, no gross changes in the toxin conformation (with smaller toxin flexibility on α7 nAChR), while slow WTX binding to nAChR is associated with slow irreversible rearrangement both of the tip of the toxin loop II and of the binding pocket residues locking finally the toxin molecule. Computer modeling showed that the additional disulfide in the loop I is not directly involved in receptor binding of WTX and candoxin, but it stabilizes the structure of loop I which plays an important role in toxin delivery to the binding site. In summary, computer modeling visualized possible modes of binding for those weak toxins which interact with the nAChR, provided no solutions for those weak toxins whose targets are not the nAChRs, and demonstrated that the additional disulfide in loop I cannot be a sound criteria for joining all weak toxins into one group; the conclusion about the diversity of weak toxins made from computer modeling is in accord with the earlier phylogenetic analysis.