INTERACTIONS OF TOXINS WITH CHOLINERGIC PROTEINS

Abraham O. Samson¹, Tali Scherf², Jordan H. Chill¹, Miriam Eisenstein², Jacob Anglister¹

¹Department of Structural Biology and ²Chemical Services

Weizmann Institute of Science

76100 Rehovot, Israel

e-mail: Jacob.Anglister@weizmann.ac.il

Abstract

The structure of a peptide corresponding to residues 182–202 of the acetylcholine receptor α1-subunit in complex with α-bungarotoxin was solved using NMR spectroscopy. The peptide contains the complete sequence of the major determinant of AChR involved in α-bungarotoxin binding. One face of the long β-hairpin formed by the AChR peptide consists of exposed non-conserved residues, which interact extensively with the toxin. Mutations of these receptor residues confer resistance to the toxin. Conserved AChR residues form the opposite face of the β-hairpin, which creates the inner and partially hidden pocket for acetylcholine. An NMR-derived model for the receptor complex with two α-bungarotoxin molecules shows that this pocket is occupied by the conserved α-neurotoxin residue R36, which forms cation-π interactions with both ^αW149 and ^γW55/^δW57 of the receptor and mimics acetylcholine. Fasciculin and α-bungarotoxin exhibit sequential similarity and overlapping binding regions but differ in their interacting residues. The complex between acetylcholine esterase and fasciculin retains some enzymatic activity, because unlike α-bungarotoxin, fasciculin does not occupy the binding site.

Introduction

Neurotoxins are classic venom component affecting neuronal transmission. Despite their highly conserved structural motifs, the toxins of this family differ in their target and mode of action. α-bungarotoxin (α-BTX) is a 74 amino-acid, 8 kDa α-neurotoxin derived from the venom of the snake Bungarus Multicinctus. It binds to the postsynaptic muscle-acetylcholine receptor (AChR) with a K_D of 10^-11 (1), competitively inhibiting acetylcholine (ACh) binding, thereby preventing the depolarizing action on postsynaptic membranes and blocking neuromuscular transmission. Peptides corresponding to linear segments of the AChR bind α-BTX with high affinity (2-7). The major determinant involved in toxin binding was mapped to the segment ^α1W184-^α1D200 which forms a β-hairpin (8).

Fasciculins (FAS1 and FAS2) are 61-residue, 7 kDa polypeptides originating from mamba venom. They are a powerful reversible inhibitors of acetylcholine esterase (AChE; EC 3.1.1.7) causing a build-up of ACh in the neuromuscular junction. Fasciculins selectively inhibit mammalian and electric fish AChE with pico- to nanomolar affinity. FAS2 exhibits higher affinity to AChE than FAS1 due to an asparagine residue instead of a tyrosine residue at sequence position 47. (α-BTX, FAS and AChR residues are designated by a superscript B, F, α1, α7, β or δ (i.e ^α1X) respectively, before the one letter amino-acid code indicating the subunit type and the position in sequence.)

Complexes of αAChR peptide analogues with α-BTX have also been the subject of structural studies. Harel et al., Scherf et al. and Scarselli et al. determined the three-dimensional structure of library derived peptides and their modified variants in complex with α-BTX (9-12). These library based peptides are homologous to neuronal (α7) AChR with no insertion or deletion in the sequence and their two structurally important prolines are located at the same position. Moise at al. solved the solution structure of a complex between α-BTX and a neuronal AChR peptide (13). Basus et al. established the structure of a short segment of the muscle AChR in complex with α-BTX (14). Yet, none of these studies describes the structure of the major muscle AChR determinant involved in toxin binding, namely the segment ^α1W184-D200, and the structure of the whole AChR has not been solved.

The crystal structure of a snail ACh binding protein (AChBP) which buffers ACh levels within the synapse was determined (15). This homopentamer shares 27% sequence identity with the α7 subunit of AChR. Superposition of the toxin bound high-affinity peptide HAP on the analogous region of the AChBP located the binding site for α-neurotoxins at the outer perimeter of the AChBP at the interface between two identical subunits (11).

The three-dimensional structure of the complex between AChE and FAS was solved concurrently in the laboratories of Sussman (16) and Marchot (17). The high affinity of FAS for AChE is due to a remarkable surface complementarity, involving a large contact area (2000Ų). Loop II of FAS inserts into the gorge of AChE, thereby sterically occluding substrate access. Loop I and the C-terminus are in close contact with the enzyme.

In the present study, we determined the solution structure of a complex between α-BTX and a peptide corresponding to the segment ^α1R182-^α1T202 containing the entire major ligand binding domain of Torpedo α1. This structure correlates the observed changes in toxin binding with the naturally occuring mutations of aAChR in different species. Using our NMR structure of α-BTX/α1^182-202 complex, we constructed a homology-based model of the extracellular domain of the AChR, AChR-EC, in complex with two toxin molecules. This model provides an explanation for the mechanism of AChR inhibition by snake a-neurotoxins. Finally, a comparative analysis of the structurally related a-BTX and FAS is presented, shedding light on their distinctive mode of actions.

Results & Discussion

Structure Determination

The structure of the α-BTX/^α1182-202 complex was determined based on a total of 1673 distance constraints including 104 between the peptide and the toxin and 118 torsion angle constraints. Figure 1A shows the backbone superposition of 28 lowest energy structures of the complex that satisfy the experimental restraints with no NOE violations larger than 0.4 Å and no torsion angle violations exceeding 5°. The overall rmsd values are of 0.84 Å and 1.45 Å for the backbone and heavy atoms, respectively (excluding peptide terminal residues ^α1R182-^α1G183 and ^α1I201-^α1T202).

Structure of the bound α-BTX/α1^182-202 complex

As shown in Figure 1B, the overall structure of α-BTX consists of three long fingers and a C-terminal tail. Finger I forms a β-hairpin with two antiparallel β-strands consisting of residues ^BV2-^BT6 and ^BI11-^BT15. Finger II consists of two antiparallel β-strands, ^BL22-^BD30 and ^BG37-^BA45. Residues ^BE56-^BC60 of finger III form a triple-stranded antiparallel β-sheet with finger II to create the central core of the toxin. These motifs are present in many α-neurotoxins (18).

As already revealed in secondary structure determination of the bound peptide (8), α1^182-202 adopts a β-hairpin conformation, consisting of two anti-parallel β-strands formed by residues ^α1H186-^α1T191 and ^α1Y198-^α1D200 and a six residue connecting loop made of ^α1C192-^α1P197 (CCPDTP) rigidified by the disulfide bond and two prolines. The first three residues of the elongated β-strand ^α1H186-^α1T191 interact with the second β-strand of α1^182-202, ^α1Y198-^α1D200, thus closing the β-hairpin, while the last three residues of the first strand, namely ^α1Y189-^α1T191, associate with the toxin residues ^BK38-^BV40, to form an intermolecular β-sheet.

^α1182-202/α-BTX Binding Interactions

Surrounded by the toxin, α1^182-202 fits snugly into the α-BTX binding site. As shown in figure 2, 12 α1^182-202 residues interact with 19 toxin residues. The sidechains of ^α1K185, ^α1W187, and ^α1Y189 interact through mostly hydrophobic interaction with residues ^BT6-^BS12 of the first finger of α-BTX. Peptide residues ^α1Y189-^α1T191 interact with residues ^BK38-^BV40 of the toxin β-sheet core through an intermolecular β-sheet involving four hydrogen bonds (figure not shown). Hydrophobic interactions between ^α1Y189 and ^BV40 on the upper side of the β-hairpin and between ^α1Y190 and ^BV39 on the lower side of the β-hairpin help stabilize the intermolecular β-sheet. The sidechains of tyrosines ^α1Y190 and ^α1Y198 on the lower side of the β-hairpin interact with ^BR36 of the toxin’s second finger, a highly conserved toxin residue found to be important for toxin binding to AChR (see below). Finally, residues ^α1Y189, ^α1T191, ^α1C192, and ^α1P194 interact through mostly hydrophobic interaction with residues ^BH68-^BQ71 at the C terminus of the toxin. Residues ^α1K185, ^α1W187, ^α1Y189, ^α1Y190, ^α1T191, ^α1C192, and ^α1P194 are the strongest contributors to the contact surface between ^α1182-202 and the toxin.

The NMR Structure of α-BTX Complex with α1^182-202 Accounts for Species-Specific Susceptibility to the Toxin

Snake neurotoxins have evolved to paralyze the snakes’ prey by inactivating muscle AChR, explaining the high affinity of long and short α-neurotoxins exhibited by muscle AChR and its α1 subunit. In Figure 2, sequences of the α1 of various species are presented together with their relative binding affinity to α-BTX. The natural preys of the snake Bungarus multicinctus are frogs and chicks, and it is therefore not surprising that α-BTX binds lethally and with the highest affinity to their α1. The Torpedo californica α1 sequence is similar to that of frogs and, therefore, exhibits similar affinities (19). On the other hand, snakes themselves and their predators such as the mongoose are naturally resistant to snake venom in general, and α-BTX in particular (20). Other species such as humans and hedgehogs, the latter being closely related to the mongoose, exhibit reduced sensitivity to α-BTX poisoning (21). Understanding the influence of a mutation on the actual binding is a powerful tool in relating α1 structure to its function.

While the upper and lower face of the β-hairpin and the backbone of the N-terminal β-strand (^α1Y189–^α1T191) are involved in toxin binding (see Figure 2A), only the lower face is involved directly in ACh binding. Resistance to snakes’ toxins can therefore be obtained by mutating residues with side chains pointing to the upper face while conserving those with side chains pointing downwards and that are crucial for ACh binding. Figure 2B indicates that mutations of residues ^α1K185, ^α1W187, ^α1Y189, ^α1P194, and ^α1P197 lead to a decrease or loss of toxin binding capability. In snakes, resistance to α-neurotoxins is conferred by the ^α1K185W, ^α1W187S, ^α1Y189N, and ^α1P194L mutations while in mongoose, resistance is obtained by ^α1W187N (putatively N-glycosylated), ^α1Y189T, ^α1P194L, and ^α1P197H mutations (21). Our structure indicates that the side chains of residues ^α1K185, ^α1W187, ^α1Y189, and ^α1P194 point to the upper side of the β-hairpin and interact extensively with α-BTX. The aforementioned mutations obviate the favorable interactions with the toxin and abolish its binding. Figure 2B also indicates that mutations of residues ^α1D195 and ^α1T196 do not significantly alter the AChR affinity to the toxin. In susceptible species such as frogs, ^α1T196 is replaced by a lysine, whereas in cats, ^α1D195 is replaced by threonine. Interestingly, T₁ relaxation time in the rotating frame (T_1ρ) and rmsd values of residues ^α1D195 and ^α1T196 suggest they are more flexible than other residues in the binding determinant (Samson et al., unpublished data). Our findings suggest that these residues are solvent exposed in α1^182-202 and do not contribute to α-BTX binding.

Modeling

Since AChBP residues 179–194 (KKNSVTYSCCPEAYEDV), which are analogous to the α-BTX bound α1^182-202, were found to adopt a β-hairpin conformation, in which Ser¹⁸⁶-Cys¹⁸⁷ form a turn (15) we opted to construct a NMR-derived model of the complex between AChR and α-BTX based on the AChBP. The backbone superposition of AChR segments ^α1K185-^α1Y190 and ^α1Y198-^α1L199 over that of the corresponding AChBP region yielded an rmsd of 1.4 Å (Figure 2B), a deviation originating mostly from the one residue insertion ^α1P194 in the AChR sequence.

The AChR model was build by replacing the sidechains of the AChBP with those of AChR as dictated by the sequence alignment. The β-hairpin (^α1K185-^α1D200, shown in light color in figure 3A) in the NMR structure of the α-BTX/α1^182-202 complex was superimposed on the corresponding β-hairpin in the AChR-EC model. The β-hairpin in the AChR model was then exchanged with the NMR structure of the entire α-BTX/α1^182-202 complex. The exchange with the whole α-BTX/α1^182-202 complex automatically dictated the position of the toxin relative to the receptor, thus generating an NMR-derived model for the α-BTX/AChR-EC complex (Figure 3). A steric clash observed between the sidechain of ^BS34 and the receptor was resolved by energy minimization, allowing movement of only three residues (^BC33-^BS35).

α-BTX forms an angle of approximately 35° with the plane of the pentameric ring of AChR and a 37° angle with the tangent to the ring (Figure 3). In contrast, the superimposed model of Harel et al. located α-BTX in the plane of the pentameric ring and perpendicular to the tangent to the AChBP ring (11). The different angular orientation of α-BTX in the AChR model dramatically increases its contact area with the receptor by a factor of ~2.5. Upon binding AChR, 37% (1869 Ų) and 34% (1745 Å ²) of the toxin surface (5013 Å ²) become buried at the α1γ and α1δ interfaces, respectively. The buried surface area of the toxin is unusually large in comparison to other protein-protein complexes and explains the very tight binding between α-BTX and AChR. The additional contact area of 124 Å ² is consistent with the higher affinity of the α1δ binding site. At the high-affinity binding site, 791 Ų of the α1 subunit and 870 Ų of the δ subunits are buried upon α-BTX binding. At the lower affinity binding site, 784 Ų of the α1 subunit and 766 Ų of the γ subunit are buried upon α-BTX binding.

Toxin Residue R36 Occupies the ACh Binding Site

The most striking feature of the NMR-derived model of the AChR/α-BTX complex is the ACh binding site occupied by ^BR36 (Figures 4A and 4B), which mimics ACh (Figure 4C). The majority of the receptor residues interacting with ^BR36 are from the α1 subunit. The positively charged guanidinium group of ^BR36 forms cation-π interactions with ^α1W149, ^δW57 (^γW55 in the γ subunit), and possibly with ^α1Y93 in agreement with previous predictions (22). In addition, a hydrogen bond is formed between the guanido group of ^BR36 and the carbonyl oxygen of ^α1W149. ^α1Y190, ^α1Y198, and ^δL121 (^γL119 in the γ subunit) interact with the methylenes of ^BR36 sidechain. ^α1C192 and ^α1C193 are close to the carbonyl group of ^BR36 and the methylene of ^BG37. Notably, the orientation of ^BR36 in the AChR/α-BTX model is dictated by the interactions with ^α1Y190 and ^α1Y198 as observed by NMR and was not changed in the modeling process.

^BR36 Is Invariant in α-Neurotoxins

Sequence alignment of several long and short α-neurotoxins was performed using ClustalW and displayed a high sequence identity (35%–65%) as well as five invariant cystine bridges (Figure 5). The alignment revealed that the arginine at the tip of the second finger, ^BR36, and ^BG37 are invariant (Figure 5). As mentioned earlier, ^BR36 occupies the ACh binding site on the receptor, while the small and flexible ^BG37 enables optimal fit of ^BR36 in the ACh binding pocket. These findings are in excellent agreement with mutagenesis results that show that a mutation of R33 of NmmI (homologous to ^BR36, see Figure 5) results in four orders of magnitude decrease in the affinity of the toxin to AChR (Osaka et al., 2000). In addition to ^BR36 and ^BG37, residues ^BW28 and ^BP49 were the only invariant residues excluding the cysteines. Remarkably, ^BW28 interacts extensively with the γ and δ subunits.

Fasciculin vs. α-Bungarotoxin

Despite their structural similarity including 43% sequence homology, a common three finger fold assembled by 5 β-strands, 4 conserved disulfide bridges, and a backbone rms deviation of 2.8Å, FAS and α-BTX do not share the same cholinergic target. The difference in fraction of buried surface upon complexation was calculated for the toxins using InsightII.

Upon binding to the receptor the second finger of α-BTX penetrates deeply into the interface between the αδ and the αγ subunits where its residues ^BK26-^BE41 interacts extensively with the subunits. Toxin residues ^BT5-^BS12 on finger I and ^BH68-^BQ71 on the C-terminus interact with the α subunit of the AChR. Finger III of the toxin interacts with the γ and δ subunits. Upon complexation with AChE, the second finger of FAS, comprising residues ^FR27-^FR37, inserts into the AChE gorge, while finger I residues ^FY4-^FA13, finger III residue ^FN47, and the C-terminal residue ^FY61 form close contacts with the AChE. FAS and α-BTX alignment was performed using ClustalW and is presented in figure 6 together with the change in fraction of buried surface per residue upon complex formation. The data show that the amino-acids located at analogous positions in the toxin sequence form the contacts with AChR and AChE respectively. These contact residues, at the tip of the three fingers and the C-terminus, are not conserved because they interact differently with their cholinergic target. The invariant α-neurotoxin residue R36 is not present in FAS because a proline residue is necessary for AChE binding. Preserved residues are encountered at positions not involved in contact formation and stem from a common precursor. The idea that FAS and α-BTX binding regions overlap but present different interacting residues is further supported by a study of Ricciardi et al. (23) in which a recombinant chimera of an AChR- and an AChE-targeted toxin was produced. Mutation of the interacting residues of a toxin-α directed against AChR with those of FAS2 aimed against AChE resulted in a chimera specific against the AChE.

FAS inhibition is achieved by capping the entrance to the AChE gorge and sterically blocking the entry of the ligand. Surprisingly, some catalytic activity of the FAS bound AChE has been reported (24). The catalytic site of the enzyme is not sealed, and it has been proposed that substrate access is permitted through a side door (25) and a back door (26) opening. In contrast, the critical role suggested in our study for ^BR36 is in perfect agreement with the complete loss of activity upon ligand binding. It is not only that the α-BTX second finger serves as a mechanical lid, preventing ACh from entering and leaving the binding pocket. Rather, the α-neurotoxin conserved R36 mimics ACh in its binding pocket and forms cation- interactions with both ^αW149 and ^γW55/^δW57 of the AChR thereby hindering ligand binding (27).

References

1. Stroud, R. M., McCarthy, M. P., and Shuster, M. (1990) Biochemistry 29, 11009-11023

2. Wilson, P. T., Hawrot, E., and Lentz, T. L. (1988) Mol Pharmacol 34, 643-650

3. Wilson, P. T., and Lentz, T. L. (1988) Biochemistry 27, 6667-6674

4. Ralston, S., Sarin, V., Thanh, H. L., Rivier, J., Fox, J. L., and Lindstrom, J. (1987) Biochemistry 26, 3261-3266

5. Conti-Tronconi, B. M., Tang, F., Diethelm, B. M., Spencer, S. R., Reinhardt-Maelicke, S., and Maelicke, A. (1990) Biochemistry 29, 6221-6230

6. Aronheim, A., Eshel, Y., Mosckovitz, R., and Gershoni, J. M. (1988) J Biol Chem 263, 9933-9937

7. Neumann, D., Barchan, D., Safran, A., Gershoni, J. M., and Fuchs, S. (1986) Proc Natl Acad Sci U S A 83, 3008-3011

8. Samson, A. O., Chill, J. H., Rodriguez, E., Scherf, T., and Anglister, J. (2001) Biochemistry 40, 5464-5473.

9. Scherf, T., Balass, M., Fuchs, S., Katchalski-Katzir, E., and Anglister, J. (1997) Proc Natl Acad Sci U S A 94, 6059-6064

10. Scherf, T., Kasher, R., Balass, M., Fridkin, M., Fuchs, S., and Katchalski-Katzir, E. (2001) Proc Natl Acad Sci U S A 98, 6629-6634.

11. Harel, M., Kasher, R., Nicolas, A., Guss, J. M., Balass, M., Fridkin, M., Smit, A. B., Brejc, K., Sixma, T. K., Katchalski-Katzir, E., Sussman, J. L., and Fuchs, S. (2001) Neuron 32, 265-275.

12. Scarselli, M., Spiga, O., Ciutti, A., Bernini, A., Bracci, L., Lelli, B., Lozzi, L., Calamandrei, D., Di Maro, D., Klein, S., and Niccolai, N. (2002) Biochemistry 41, 1457-1463.

13. Moise, L., Piserchio, A., Basus, V. J., and Hawrot, E. (2002) J Biol Chem 277, 12406-12417.

14. Basus, V. J., Song, G., and Hawrot, E. (1993) Biochemistry 32, 12290-12298

15. Brejc, K., van Dijk, W. J., Klaassen, R. V., Schuurmans, M., van Der Oost, J., Smit, A. B., and Sixma, T. K. (2001) Nature 411, 269-276.

16. Harel, M., Kleywegt, G. J., Ravelli, R. B., Silman, I., and Sussman, J. L. (1995) Structure 3, 1355-1366.

17. Bourne, Y., Taylor, P., and Marchot, P. (1995) Cell 83, 503-512.

18. Tsetlin, V. (1999) Eur J Biochem 264, 281-286.

19. Ohana, B., and Gershoni, J. M. (1990) Biochemistry 29, 6409-6415

20. Barchan, D., Kachalsky, S., Neumann, D., Vogel, Z., Ovadia, M., Kochva, E., and Fuchs, S. (1992) Proc Natl Acad Sci U S A 89, 7717-7721

21. Barchan, D., Ovadia, M., Kochva, E., and Fuchs, S. (1995) Biochemistry 34, 9172-9176.

22. Zhong, W., Gallivan, J. P., Zhang, Y., Li, L., Lester, H. A., and Dougherty, D. A. (1998) Proc Natl Acad Sci U S A 95, 12088-12093

23. Ricciardi, A., le Du, M. H., Khayati, M., Dajas, F., Boulain, J. C., Menez, A., and Ducancel, F. (2000) J Biol Chem 275, 18302-18310.

24. Radic, Z., Duran, R., Vellom, D. C., Li, Y., Cervenansky, C., and Taylor, P. (1994) J Biol Chem 269, 11233-11239.

25. Tara, S., Straatsma, T. P., and McCammon, J. A. (1999) Biopolymers 50, 35-43.

26. Gilson, M. K., Straatsma, T. P., McCammon, J. A., Ripoll, D. R., Faerman, C. H., Axelsen, P. H., Silman, I., and Sussman, J. L. (1994) Science 263, 1276-1278.

27. Samson, A., Scherf, T., Eisenstein, M., Chill, J., and Anglister, J. (2002) Neuron 35, 319-332.