TETANUS TOXIN H

Paul Emsley, Neil Fairweather, Ian G. Charles, Neil W. Isaacs

Deptment of Chemistry, University of Glasgow Hillhead, Glasgow, G12 8QQ
Imperial College, University of London
The Cruciform Project, Imperial College, University of London

Keywords: Crystal structure, TeNT, Fragment-C, ganglioside binding

Tetanus toxin (TeNT) and botulinum toxin (BoNT) are amongst the most potent toxins known. They are produced by Closdridial bacteria (C. tetani and C. botulinum, respectively) and the purpose is to kill the host via paralysis and suffocation.The toxins are structurally related (about 50% sequence identity) and functionally related. However, TeNT targets the neurons of the spinal cord via trans-synaptic migration, whereas BoNT targets secondary neurons.

The mechanism of these toxins consists of 4 stages, cell binding, vesicular internalisation, cytoplasmic translocation and finally proteolytic cleavage of the substrate by the L-chain. It is in order to understand these functions that we have undetaken the high resolution structure of Tetanus toxin H.

TeNT is synthesised as a single polypeptide chain which undergoes cleavage to produce a mature toxin consisting of the N-terminal 50kD fragment linked via a disulphide bond to to 100kD C-terminal fragment (H chain). The 50kD fragment of the C-terminal portionn of the H chain (know as H, or fragment-C) is responsible for ganglioside binding, which is essential for binding of the toxin to neuronal cells.

Crystals of H were grown in conditions similar to those published previously with the modification of the addition of PEG 4K and 1% MPD. Cryo-cooled data collection was performed at Daresbury stations 7.2, 9.5 and 9.6:

• 3.0A UAc data,
• 3.0A PtCl data
• 2.5A HgAc + CHHgCl data
• 3.0A HgAc + CHHgCl (different soak-time and wavelength)
• 1.8A data (Daresbury 7.2)

The 2.5A Hg dataset (compared to the 2.3A Native) provided the position of the leading 3 Hg peaks. The other 11 sites were found from difference Fouriers and similar maps from SHARP. Density modification (DM) gives provided a map in which most of the model was traced. A few remaining loops were tracing with the aid of a picture of a CA-representation provided in .

The high resolution native showed considerable lack of isomorphism compared to the lower (2.3A) data (scaling R-factors typically 20-30%). Therefore, AMORE was then used to reposition the molecule in the unit cell of the high resolution data. The refinement proceeded conventionally, using REFMAC and ARP finding 3 glycerol molecules, 400 water molecules and geometry judged acceptable for a typical 1.8Astructure as judged from the output of PROCHECK.

Structural conclusions: Tetanus toxin H consists of two domain, the N-terminal domain is a -sandwich and the C-terminal domain is a -trefoil. Then -trefoil contains the ganglioside binding sites. Several glycerol molecules have been observed in the crystal structure and as has been seen previously may indicate the carbohydrate binding site.