32_Mordvintsev 10/11/06 3:09 PM Page 71

Journal of Molecular Copyright © 2006 Humana Press Inc. All rights of any nature whatsoever are reserved. ISSN0895-8696/06/30:71–72/$30.00 JMN (Online)ISSN 1559-1166 DOI 10.1385/JMN/30:1-2:71

ORIGINAL ARTICLE

A Model for Short F- Bound to Nicotinic Receptor From Torpedo californica

Dmitry Y. Mordvintsev,*,1 Yakov L. Polyak,1 Dmitry A. Kuzmine,1 Olga V. Levtsova,2 Yegor V. Tourleigh,2 and Igor E. Kasheverov1

1Shemyakin-Ovchinnikov Institute of Bioorganic Chemistry RAS, 117997, GSP-7, Moscow, Russia; and 2M. V. Lomonosov Moscow State University, Biology Faculty, Department of Biophysics, 111992, Moscow, Russia

Introduction Building of the Models 1 Short- and long-chain F- from snake The basis of this model is a set of H-NMR struc- venoms are potent blockers of nicotinic acetylcholine tures of NTII and a wealth of labeling and muta- receptors (nAChRs). Short F-neurotoxins consist of genesis data available for this and other homologous 60–62 residues and include 4 disulfide short-chainF-neurotoxins. When building the model bridges, whereas long F-neurotoxins have 66–75 ofT. californicanAChR extracellular domain, not only residues and 5 disulfides. The spatial structure of the cryoelectron microscopy data for the T. marmorata these toxins is built by three loops, I–III “fingers,” receptor (free of any ligands) were taken into account, confined by four disulfide bridges; the fifth disulfide but also the data for AChBP complexes with ago- of long-chain F-neurotoxins is situated close to the nists and antagonist. Comparative modeling was tip of central loop II. An accurate knowledge of the done with Modeller program, and toxin-receptor mode of F-neurotoxin–nAChR interaction is impor- interactions were simulated with the biochemically tant for rational design of new nAChR agonists and managed approach under Haddock. The data on antagonists for medical purposes. Ideas on the topo- labeling with azidobenzoyl derivatives of NTII have graphy of toxin–nAChR complexes were based until been used for this purpose. Subsequent molecular recently on nAChR interactions with selectively dynamic simulations were compared with data on labeled F-neurotoxins, mutations in toxins, nAChR, the mobility and accessibility of spin and fluores- or both. Recently, crystal structures have been solved cence labels in free and bound NTII. Final model for the Torpedo marmorata nAChR (4Å[Unwin, 2005]) structures were analyzed. and for the acetylcholine-binding protein (AChBP) Model Analysis complexed with mollusk F-conotoxin (2.4 Å[Celie et al., 2005]) or F-cobratoxin, long-chain F-neurotoxin NTII was located at approx 25–30 Åfrom the mem- (4 Å [Bourne et al., 2005]). However, there were no brane surface. Toxin molecular axis, defined by the angstrom-resolution models for complexes of short- direction of the G3-sheet, lies at an ~80° angle rela- chain F-neurotoxins. Here, we report the model of tive to the median axis from the center of the nAChR the Torpedo californica nAChR extracellular domain extracellular domain ring and at a small angle to the complexed to a short-chain F-neurotoxin II (NTII) cylinder wall, practically perpendicular to the mem- from Naja oxiana cobra venom. brane surface. About one-third of toxin loop II is

*Author to whom all correspondence and reprint requests should be addressed. E-mail: [email protected]

Journal of Molecular Neuroscience 71 Volume 30, 2006 32_Mordvintsev 10/11/06 3:09 PM Page 72

72 Mordvintsev et al.

plunged into the ligand-binding pocket, located at residues in the binding pocket not observed for the interface between two nAChR subunits, whereas F-conotoxin binding. Being in general very similar, loops I and III contact the receptor residues by their binding modes of short- and long-chain F-neuro- tips only. The concave side of loops II and III is very toxins differ considerably in the manner of loop II close to the protruding C-loop of the F-subunit, entry into nAChR. whereas loop I, the amino terminus, and convex side of loop III are solvent accessible. Loop III also con- References tacts the F-loop of the adjacent subunit. The toxin Bourne Y., Talley T. T., Hansen S. B., Taylor P., and Mar- structure undergoes some changes during the final chot P. (2005) Crystal structure of Cbtx-AChBP com- complex formation (for rmsd 1.45 in 15–25 ps), which plex reveals essential interactions between snake correlates with earlier NMR data. This position of alpha-neurotoxins and nicotinic receptors. EMBO J. 24, NTII is in good agreement with the available bio- 1–11. chemistry and data. Celie P. H., Kasheverov I. E., Mordvintsev D. Y., Hogg R. The binding process was shown to be dependent C., van Nierop P., van Elk R., et al. (2005) Crystal struc- ture of nicotinic acetylcholine receptor homolog mostly on spontaneous outward movement of the AChBP in complex with an F-conotoxin PnIA variant. C-loop found earlier in the AChBP complexes with Nat. Struct. Mol. Biol. 12, 582–588. F-cobratoxin and F-conotoxin. Among common Unwin N. (2005) Refined structure of the nicotinic acetyl- features in binding of short- and long-chain choline receptor at 4Å resolution. J. Mol. Biol. 346, F-neurotoxins is the rearrangement of aromatic 967–989.

Journal of Molecular Neuroscience Volume 30, 2006