Báo cáo khoa học: Solution structure of long neurotoxin NTX-1 from the venom of Naja naja oxiana by 2D-NMR spectroscopy

doi:10.1111/j.1432-1033.2004.04465.x

Eur. J. Biochem. 271, 4950–4957 (2004) (cid:1) FEBS 2004

Solution structure of long neurotoxin NTX-1 from the venom of Najanajaoxianaby 2D-NMR spectroscopy

Mehdi Talebzadeh-Farooji1, Mehriar Amininasab1, Maryam M. Elmi1, Hossein Naderi-Manesh2 and Mohammad N. Sarbolouki1 1Institute of Biochemistry & Biophysics, University of Tehran, Iran; 2Faculty of Science, Tarbiat Modares University, Tehran, Iran

0.23 and 1 A˚ for residues involved in b-sheet regions, respectively. The overall fold in the NMR structure is similar to that of the X-ray crystallography, although some differ- ences exist in loop I and the tip of loop II. The most func- tionally important residues are located at the tip of loop II and it appears that the mobility and the local structure in this region modulate the binding of NTX-1 and other long neurotoxins to the nicotinic acetylcholine receptor.

long neurotoxins; neurotoxin-1;

Keywords: 2D-NMR; solution structure; three-finger peptide.

The NMR solution structures of NTX-1 (PDB code 1W6B and BMRB 6288), a long neurotoxin isolated from the venom of Naja naja oxiana, and the molecular dynamics simulation of these structures are reported. Calculations are based on 1114 NOEs, 19 hydrogen bonds, 19 dihedral angle restraints and secondary chemical shifts derived from 1H to 13C HSQC spectrum. Similar to other long neurotoxins, the three-finger like structure shows a double and a triple stranded b-sheet as well as some flexible regions, particularly at the tip of loop II and the C-terminal tail. The solution NMR and molecular dynamics simulated structures are in good agreement with root mean square deviation values of

His67 play an important role in receptor binding and hence in the toxicity of neurotoxins [1,2,6,8].

Neurotoxins belonging to Elapidae family are divided [1–3] into two groups of short (60–62 amino acids in length with four disulfide bridges) and long neurotoxins (66–74 amino acids with five disulfide bridges). They cause postsynaptic blockage of the nicotinic acetylcholine receptor (AChR) [4,5]. Despite their functional similarity, these peptides have differences in their amino acid sequences such as the deletion/insertion of some residues and the presence of longer C-terminal tails in long neurotoxins (Fig. 1). It is also known that the association of long neurotoxins with their receptor, and also their dissociation, occurs more slowly than short neurotoxins [1].

Already the crystal and solution structures of some long neurotoxins have been reported [9–16]. They usually have a three-finger like structure emerging from a core (globular head) with three loops wherein loop II plays a key role in binding to AChR. Most of the functionally invariant residues are located here [17]. Basus et al. studied the complex between a-bungarotoxin (from Bungarus multi- cinctus) and the nicotinic receptor peptide by 2D-NMR spectroscopy [6] and reported some differences between the bound and unbound conformations of this neurotoxin particularly at the tip portion of loop II. Comparison of the crystal and solution structures of a-cobratoxin (a long neurotoxin from Naja naja siamensis, which has 63% homology with NTX-1) revealed that the solution structure is more flexible in the tip of loop II [14].

The binding site of neurotoxins is located at the AChR a-subunit between residues 173 and 204, and probably encompasses the agonist-binding site [6]. Because the binding affinity of neurotoxins to the target receptor is higher than that of acetylcholine, they can completely prevent acetylcholine binding [7]. Previous reports have suggested that residues Trp27, Lys25, Arg35, Lys37 and

Here we report the elucidation of solution structure of neurotoxin-1 (NTX-1; P01382, from Naja naja oxiana) via two-dimensional 1H-NMR spectroscopy. NTX-1, a long is one of the lethal neurotoxin with 73 amino acids, components of Naja naja oxiana venom. This peptide is different from other neurotoxins as it lacks a phenylalanine residue and has a lower net positive charge [18]. The NMR structure of NTX-1 is compared with the X-ray structure of PDB entry 1NTN [12] whose amino acid sequence differs in two positions: deletion of Pro9, and substitution of Asp63 with Asn.

Materials and methods

Sample preparation and purification

Central Asian cobra (Naja naja oxiana) snakes were collected, milked and the pooled venom lyophilized and

Correspondence to H. Naderi-Manesh, Faculty of Science, Trabiat Modares University, Tehran, Iran, PO Box 14115-175. Fax/Tel.: +98 218 009730, E-mail: naderman@modares.ac.ir and M. N. Sarbolouki, Institute of Biochemistry & Biophysics, University of Tehran, Tehran, Iran, PO Box 13145-1384. Fax: +98 216 404680, Tel.: +98 216 491267, E-mail: sarbol@ibb.ut.ac.ir Abbreviations: AChR, nicotinic acetylcholine receptor; a-BTX, a-bungarotoxin; a-CTX, a-cobratoxin; ESI, electron spray ionization; HSQC, heteronuclear single quantum coherence; MD, molecular dynamics; RMSD, root mean square deviation; RP-HPLC, reverse phase high performance liquid chromatography; TPPI, time proportional phase incrementation. (Received 26 June 2004, revised 13 October 2004, accepted 27 October 2004)

NTX-1 solution structure from N. naja oxiana (Eur. J. Biochem. 271) 4951

(cid:1) FEBS 2004

Secondary structure restraints These were predicted from the observed 13Ca, 13Cb and Ha resonances and the use of the chemical shift indexing [25] method along with the TALOS program [26].

stored at )20 (cid:2)C. Purification of NTX-1 was performed by the successive application of gel filtration chromatography and RP-HPLC methods. The purity, molecular mass, and correspondence with SWISS-PROT code P01382 of NTX-1 were confirmed by electron spray ionization/mass spectro- metry (ESI-MS).

Structure calculations

NMR spectroscopy

Three-dimensional structure calculations were carried out on a PC workstation using CNS [27] and the standard protocol of ARIA 1.1 program [28] under Red Hat Linux 8.0. A simulated annealing protocol in torsion angle space was used starting from an extended conformation. The protocol was implemented in four stages: (a) high temperature simulated annealing at 10 000 K; (b) a first cooling phase from 10 000 K to 1000 K in 5000 steps; (c) a second cooling phase from 1000 K to 50 K in 4000 steps; and (d) 4000 steps refinement and energy minimization. The time step for integration was set to 0.003 ps. One hundred structures were generated in each iteration, and the 20 lowest energy structures in the final iteration were evaluated through PROCHECK [29] and used for further analysis. The structural models were visualized with the program MOLMOL [30].

Molecular dynamics simulations

NTX-1 was dissolved in D2O/H2O (10 : 90, v/v) pH 3.2 or 99.9% (v/v) D2O to the final concentration of 4 mM. All NMR spectra were recorded on a Bruker DRX-500 spectrometer. Water suppression was achieved by applying WATERGATE pulse sequence. Two-dimensional 1H- NMR spectra were acquired at 20 and 37 (cid:2)C to overcome possible ambiguities in assignments. The spectral width was adjusted to 12.061 p.p.m. (6009.62 Hz) in both dimensions. The carrier frequency was set with respect to the center of residual water signal and 2,2-dimethyl-2-silapentane-5- sulfonate was used as an internal reference. The time proportional phase incrementation (TPPI) and States-TPPI methods for quadrature detection were used for NOESY [19,20]/TOCSY [19–21] and DQF-COSY [19,20,22] spectra, respectively. NOESY and TOCSY spectra were acquired at 50, 100, 150 and 200 ms mixing times with 56 scans per each t1 increment, and 50, 75 and 150 ms mixing times with 48 scans per each t1 increment, respectively. The DQF-COSY spectrum was acquired in phase sensitive mode with 184 scans per each t1 increment. The spectra were collected using 512 and 2048 complex points in t1 and t2 dimensions. Zero filling was performed in t1 dimension to obtain a final matrix of 1024 · 2048 data points.

A natural abundance 1H-13C HSQC spectrum was acquired at 20 (cid:2)C with 200 scans per each t1 increment. The spectral widths in F1 and F2 dimensions were adjus- ted to 165.66 p.p.m. (20 833.33 Hz) and 13.333 p.p.m. (6666.67 Hz), respectively. All spectra were processed by XWIN-NMR software and analyzed with XEASY [23].

Molecular dynamics (MD) simulations were performed on a PC workstation using the SANDER module of AMBER 7.0 [31] under Red Hat Linux 8.0. All of the calculations were carried out in explicit water with a solvent box whose edges were 8 A˚ apart from the closest protein atoms. The five best ARIA structures were first energy minimized in 5000 steps, subjected to 100 ps (2 fs time steps) of MD at constant volume and then gradually heating up from 200 K to 300 K, followed by 100 ps (2 fs time steps and pressure relaxation time of 2 ps) of MD at 300 K and constant pressure. Finally, 500 ps MD simulation with 2 fs time steps and 2 ps pressure relaxation time at 300 K were carried out.

Distance restraints

Fig. 1. Multiple sequence alignment of NTX-1. NTX-1 is aligned with a number of typical long and short neurotoxins, whose three- dimensional structures have been discovered.

Results and Discussion

Resonance assignment

Measuring the cross-peak intensities from the 150 ms mixing time NOESY spectrum resulted in interproton restraints for NTX-1. The results of hydrogen–deuterium exchange experiments along with preliminary NOE based calculated structures were used to deduce hydrogen bond restraints which were assigned to 2.8 ± 0.5 A˚ for rN-O and 1.8 ± 0.5 A˚ for rH–O.

NMR assignments were carried out according to standard methods [32]. Identification of spin systems was made using the DQF-COSY and TOCSY spectra recorded at 20 (cid:2)C. For assignment of the Ha cross-peaks in the region of water suppression, the spectra obtained at a higher temperature, 37 (cid:2)C, were used wherein the water signal was shifted relative to Ha resonances. The results of the spin system identification were confirmed by the 1H-13C HSQC spec- trum. The chemical shifts table of 13C-NMR for Ca and Cb and 1H-NMR has been deposited in BioMagResBank (BMRB), code 6288 (Table S1). To perform the sequential

/ Angle restraints 3JNHa constants were determined from 2D NOESY spec- trum using the program INFIT [24] and / dihedral angles were restrained to the ranges of )120(cid:2) ± 30(cid:2) for 3JNHa > 8 Hz and )60(cid:2) ± 30(cid:2) for 3JNHa < 6 Hz.

4952 M. Talebzadeh-Farooji et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

38–44 of loop II along with segment 55–59 of loop III form a triple-stranded antiparallel b-sheet held together via hydrogen bonds. Loop II has a bulky tip with an extra disulfide bridge, Cys28-Cys32, which is not present in short neurotoxins. Although the tip portion of loop II it is poor in medium and long range NOEs (Fig. 2), shows a local structure like an a-helix in the segments containing residues 32–36, as revealed by Ha32-HN35 and Ha32-HN36 NOEs.

assignment, the identified spin systems were sequentially connected by observing Ha-HN, HN-HN and Hb-HN cross- peak NOEs in 150 ms NOESY spectrum. Two successive Ala residues, Ala44 and Ala45, and also Trp27 and Trp31 were used as the starting points. The pattern of sequential NOEs for prolines showed that they are in trans confor- mation. Recognition of the secondary structural elements was carried out using long range NOEs between Ha protons, protection of HNs in deuterium exchange experi- ments and the secondary chemical shifts from HSQC spectrum. These results indicated the presence of a double- and a triple-stranded antiparallel b-sheet.

Loop III, spanning residues 45–59, has two legs, one of which participates in the third strand of the triple stranded b-sheet while the second one having no involvement in regular secondary elements shows a relatively fixed structure as revealed by the low root mean square deviation (RMSD) in this segment. A disulfide bond between Cys47 and Cys58 stabilizes this loop and a type I turn formed by residues 51–54 connects its two legs.

Structure calculations were based mainly on the NOE restraints, with its sequence distribution being shown in Fig. 2, and other restraints given in the Table 1. On the basis of the total energy, the best 20 refined structures were selected for further analysis and deposited in RCSB protein data bank, PDB code 1W6B. The superimposed structures are displayed in Fig. 3 and their geometric statistics and energetics are summarized in Table 1. All structures are in good agreement with the experimental restraints, none having NOE violations greater than 0.2 A˚ . The PROCHECK analysis of 20 best structures indicates that 97% of nonglycine and nonproline residues lay in the most favored and allowed regions of the Ramachandran plot, while only 0.3% residues are in the disallowed regions (Table 1).

Structure description

The three-dimensional structure shows a globular head with the emergence of three-finger like loops. The loops are cross- linked together by four disulfide bridges, Cys3-Cys22, Cys15-Cys43, Cys47-Cys58 and Cys59-Cy64, and are involved in a double- and a triple-stranded b-sheet, i.e. the main regular secondary structures of NTX-1. Loop I (residues 1–15) constitutes a double-stranded antiparallel b-sheet (residues 2–5 and 11–14) as well as a linker segment consisting of residues 6–10. This double-stranded b-sheet is stabilized by three hydrogen bonds.

The C-terminal tail of NTX-1 is connected to loop III via another turn-like segment containing residues 60–63 and is tethered to it by Cys59-Cys64. This region consists of two distinct segments; the first one, residues 64–68, has a more defined structure in contrast to the second, residues 69–73. It seems that the presence of the disulfide bond and the two proline residues (Pro66 and Pro68) lead to lower mobility of this segment. In addition many long range NOEs between the main chain as well as the side chains of Asn65, Pro66, His67, Pro68 and some residues in loop I and loop II (such as Tyr4, Val38, Ile39, Glu40 and Leu41), show that this region is in close contact with the other parts of the structure. Most of the calculated structures reveal a one-turn a-helix in this region as is confirmed by the observed medium range NOEs. This local structure is not observed in other long neurotoxins. On the other hand, the absence of any NOE between residues 69–73 and other regions of the peptide indicates that there is no structural involvement in this region. In contrast to the C-terminal tail, the N-terminal segment has a relatively more defined structure: both Ile1 and Thr2 are engaged in hydrogen bonding, but it seems that Cys15 is not a part of the b-sheet in loop I.

The largest

The peptide structure has a convex and a concave face (Fig. 4); the latter being instrumental in receptor binding [2]. The side chains of Leu21, Tyr23, Lys25, Ala44, Pro48 and Ile56 constitute a hydrophobic cluster on the concave face of

loop of NTX-1 (loop II) consists of residues 21–44, which is cross-linked to loop I by a disulfide bond between Cys3 and Cys22. The segment connecting loop I and loop II forms a type II b-turn by Ala16-Pro17-Gly18-Gln19 residues. Segments 21–27 and

Fig. 2. The number of NOEs per residue used in final structure calculation. Intra residue, sequential, medium range and-long range NOEs are depicted as hatched, white, grey and black filled bars, respectively.

NTX-1 solution structure from N. naja oxiana (Eur. J. Biochem. 271) 4953

(cid:1) FEBS 2004

Table 1. Structural statistics for ensemble of 20 best structures.

Parameter Value

the 500 ps of trajectory used for analysis, the structures were nearly stable and the RMSD along the trajectory was about 3 A˚ for the backbone atoms relative to the initial structure and reduced to 1 A˚ when only regular secondary elements were regarded.

Distance restraints

3JNHa Secondary chemical shifts

Intra residue (i-j ¼ 0) Sequential (|i-j| ¼ 1) Medium range (|i-j| < 5) Long range (|i-j| > 5) Hydrogen bonds Total 540 252 39 283 19 1133 Dihedral-angle restraints

19 35

In Fig. 5 parameters such as B-factor, RMSD and atomic fluctuation per residue for the X-ray and NMR structures and that obtained along the MD trajectory are shown. In this figure the free C-terminal tail (residues 69–73) has been omitted due to its high level of fluctuation. The limited experimental restraints for the tip portion of loops I, II and III (especially loop II) is consistent with the moderate level of fluctuation along the MD trajectory. As the tip of loop II is involved in binding to the target receptor it seems that flexibility of this region plays a crucial role in adopting the correct conformation. However the possession of residual structure and flexibility for the residues at the tip portion of loop II are supportive of rigid body motion [13] in the way that Gly36 serves as a hinge.

Comparison with X-ray structure

0.027 ± 0.002 0.51 ± 0.15 Mean RMS dihedral from experimental restraints NOE (A˚ ) Dihedrals ((cid:2))

0.003 ± 0.00017 0.44 ± 0.014 0.36 ± 0.013 Mean RMS deviation from idealized covalent geometry Bonds (A(cid:2)) Angles ((cid:2)) Improper ((cid:2)) Mean energy (kcalÆmol)1)

42.84 ± 7.32 380.78 ± 6.84 )607.98 ± 6.95 8.87 ± 1.04 11.016 ± 1.88 58.72 ± 3.65 )2071 ± 51.29 )2220.2 ± 50.73 ENOE Ecdih Evdw Ebond Eimproper Eangle Eelc Etotal

PROCHECK Ramachandran plot analysis for the best 20 structures Most favored region (%) Additionally allowed region (%) Generously allowed region (%) Disallowed region (%) Atomic RMS differences (A˚ )

72.6 24.7 2.4 0.3

Back bone

In Fig. 6C the crystal and the average of NMR structures are superimposed over the backbone atoms. The RMSD for the backbone atoms of the b-sheet part is 0.72 A˚ , indicating nearly identical conformation for the solution and crystal state in this region. The secondary elements of the X-ray structure can be observed in the NMR structures but the length of the b-strand in loop I in solution is somewhat shorter than that of the crystal one (Fig. 6A,B). The presence of an additional proline in position 9 in NTX-1 (P01382) makes loop I bulkier and more rigid than the corresponding loop in the X-ray structure. It is interesting to note that in contrast to the crystal structure there is no hydrogen bond network at the tip of loop II, thus residues in this region experience a fast hydrogen–deuterium exchange in D2O. The structure local a-helix show a tendency towards calculations formation (similar to the corresponding X-ray structure) whereas flexibility of the tip of loop II leads to the largest difference between the two structures in this region. For example, the side chain of Arg35 is in proximity to Trp31 in the crystal structure, whereas the solution structure shows that the aromatic side chains of Trp27 and Trp31 are close together (Fig. 6).

1.40 0.35 0.23 Residues 1–73 Residues 1–27, 38–68 b-sheets Heavy atom

the molecule, which is postulated to be involved in stability and maintenance of the whole structure [2]. Almost all of these residues are conserved in long neurotoxins (Fig. 1). The side chain of Tyr23 extends over Gly42 on the neighboring strand and therefore its HN is affected by the ring current of the aromatic group leading to an up-field resonance frequency.

The segments connecting the two legs of loop III (type I b-turn) as well as the segment intervening the loop I and II (type II b-turn) have a similar conformation in both structures. However, the regular turn elements seen at the tip portion of loop I and residues 60–63 in the crystal structure are not observed in the solution struc- ture. Comparison of crystallographic B factor and RMSD of the NMR structures reveals a satisfactory agreement between the two structures (Fig. 5), with both having comparable mobility in the same structural regions.

Molecular dynamics

Comparison with other neurotoxins

NTX-1 has a significant sequence homology with the following long neurotoxins: toxin b [33], a-BTX [34], a-cobratoxin [35] and LSIII [36] with 67%, 63%, 63%

The structures resulting from MD simulations agree satis- factorily with those of solution structure ensemble. The RMSD for the backbone atoms between the starting and the solvated-minimized structures is about 0.7 A˚ . During

Residues 1–73 Residues 1–27, 38–68 b-sheets 1.96 0.65 0.62

4954 M. Talebzadeh-Farooji et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

Fig. 3. The ensemble of 20 superimposed structures. A stereoview of the molecules is illustrated; all of the structures are fitted to the best energetic one which has been indicated as the neon. Regions involved in the formation of b-sheets are indicated in cyan.

also account for its higher flexibility [37]. On the other hand, the substitution of Ser in other long neurotoxins with Gly33 in NTX-1 (Fig. 1) at the tip portion of loop II makes this region conformationally more variable.

Binding to the nicotinic acetylcholine receptor (AChR)

and 59% identity, respectively (Fig. 1). These toxins all have 10 cysteine residues that play a major role in determining and maintaining their rigid structures. The structural superposition of these toxins reveals that their overall folding is similar. They all show a conformational variability at the tip portion of loop II, with the difference being that the length of the b-strands in NTX-1 is longer than in the others. These neurotoxins have common structural features that in many respects are analogous to short ones, although the length of loop I is longer in short neurotoxins.

It has been already shown that long neurotoxins are capable of binding to a7-neuronal as well as muscular AChR [38]. The results from mutational analysis of a-cobratoxin (a-CTX) indicate that replacement of at least eight residues (Arg33, Lys49, Asp27, Lys23, Phe29, Trp25, Arg36 and Phe65) to unrelated ones causes a remarkable decrease in its affinity towards muscular AChR [39]. These residues, mostly located at the tip portion of loop II, are conserved or conservatively substituted in the majority of long neurotoxins (Fig. 1). Also the binding role of Lys49 to muscular AChR contributes to the importance of loop III whereas similar studies have showed a trivial participation of this loop in attachment to a7-neuronal AChR [40,41]. However the role of positively charged residues in the binding of a-CTX to the muscular receptor is remarkable

The NMR structure of NTX-1 is closest to that of the these two a-cobratoxin crystal. The superposition of structures shows an RMSD value of 3.00 A˚ for all backbone atoms which reduces to 1.4 A˚ when the C-ter- minal tail and the tip of loop II are ignored. Thus there exists an almost parity between the secondary elements in the two structures. It appears that among long neurotoxins a-BTX has some unique structural features, e.g. while the side chains of Trp27, Arg33 and Asp29 in these peptides protrude from the concave face of the molecule (a common situation), the aromatic side chain of Trp27 in a-BTX occupies the convex side. The limited b-strands in a-BTX

Fig. 4. The concave (left) and convex face (right) of NTX-1. The distribution of positive, negative and hydrophobic residues are shown in blue, red and white, respectively.

NTX-1 solution structure from N. naja oxiana (Eur. J. Biochem. 271) 4955

(cid:1) FEBS 2004

does not have these two positively charged residues, shows a weaker and more reversible neurotoxic activity [36]. On the other hand Endo et al. [42] have found that the overall net charge of neurotoxins affects the toxin–receptor inter- actions. Particularly, it should be noted that NTX-1 and LSIII have a net charge of +1 in comparison with +5, +4 and +3 for a-CTX, toxin b and a-BTX, respectively, at pH 7.

However, this is not the only factor affecting the binding of neurotoxins to their receptors. As stated earlier the charac- teristic feature of all long neurotoxins, playing an important role in their function, is flexibility at the tip portion of loop II. This affects both the kinetics and thermodynamics of their interactions with the receptor [13,43,44] in a manner that lowers the free energy barrier for complex formation and increases the rate of association. Consequently, the presence of Gly33 instead of Ser at the tip of loop II in NTX-1 makes it more flexible than that of other long neurotoxins (a-BTX, a-CTX, toxin b and LSIII) and allows it to have more conformational diversity when searching for a suitable conformer in the binding cleft. On the other hand a large flexibility at this region leads to a higher entropic cost during binding of NTX-1 to its receptor, thereby lowering the binding affinity [13]. However, the rigid body motion emerging from some local structure at the tip of loop II decreases the entropic penalty upon the receptor binding. Local structures impose a restriction on freedom in this region and lower the entropy difference between the free and bound forms of the molecule [44].

and four of the above mentioned mutation-sensitive residues are cationic. Because AChR is composed of acidic subunits [2] it seems that the receptor has favorable interactions with cationic groups. Thus it may be expected that the substitution of Arg36 and Lys49 in a-CTX with Val38 and Glu51 in NTX-1, respectively, lead to a weaker interaction with muscular AChR and to a lesser extent with a7-neuronal AChR. Interestingly, LSIII, which like NTX-1

Furthermore poor experimental restraints and possibly bulkiness of the tip portion of loop II in long neurotoxins compared to short ones, as well as the finding of Bystrov et al. [45], point to the fact that this segment in the former is likely to be more flexible. This provides a possible explan- ation for the lower dissociation constants of long neuro- toxins from nicotinic acetylcholine receptor compared to short ones.

Fig. 5. The comparison of the experimental and MD results. Per residue B factor of backbone atoms of crystalline 1NTN (top), the per residue RMS difference from the mean NMR structure (middle) and per residue atomic fluctuations along the trajectory of the MD simulation for the five of the 20 best NTX-1 NMR structures (bottom) for resi- dues 1–68.

Fig. 6. Comparison of NTX-1 NMR and 1NTN X-ray structures. (A) and (B) show the ribbon representation of the averaged NTX-1 NMR and 1NTN X-ray structures, respectively. The hydrophobic side chains along with those important for receptor binding are depicted. (C) The superimposition of the averaged NTX-1 NMR (blue) and 1NTN X-ray (red) structures over the backbone atoms.

4956 M. Talebzadeh-Farooji et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

Conclusions

from Naja naja siamensis at the 2.4 A˚ resolution. J. Biol. Chem. 266, 21530–21536.

11. Dewan, J.C., Grant, G.A. & Sacchettini, J.C. (1994) Crystal structure of a-bungarotoxin at 2.3 A˚ resolution. Biochemistry 33, 13147–13154.

12. Nickitenko, A.V., Michailov, A.M., Betzel, C. & Wilson, K.S. (1993) Three-dimensional structure of neurotoxin-1 from Naja naja oxiana venom at 1.9 A˚ resolution. FEBS Lett. 320, 111–117. 13. Connolly, P.J., Stern, A.S. & Hoch, J.C. (1996) solution structure of LSIII, a long neurotoxin from venom of Laticauda semifasciata. Biochemistry 33, 418–426.

14. Le Goas, R., LaPlante, S.R., Mikou, A., Delsuc, M.A., Guittet, E., Robin, M., Charpentier, I. & Lallamand, J.Y. (1992) a-Co- bratoxin: Proton NMR assignments and solution structure. Bio- chemistry 31, 4867–4875.

It is concluded that although this group of structures has a typical three-finger shape, they have subtle differences with each other. The NMR structure reveals some minor conformational changes that might be very important in studying the mechanism of their receptor binding. The importance of cationic residues in direct binding of a-CTX to muscular as well as a7-neuronal AChRs has been demonstrated previously [39–41], and thus mutations lead- ing to the removal of the positive charge at Lys25, Arg35 and Lys37 in NTX-1 may support this suggestion. Addi- tionally, the substitution of Val38 and Glu51 with positively charged side chains (such as Arg and Lys, similar to the corresponding residues in a-CTX that are involved in receptor binding) would be interesting because of the resultant alteration in the binding affinity. Considering flexibility of loop II in receptor binding, substitution of Gly33 with Ser (a common residue in other long neurotox- ins) may lower mobility, thereby affecting the binding affinity of NTX-1.

15. Oswald, R.E., Sutcliffe, M.J., Bamberger, M., Loring, R.H., Braswell, E. & Dobson, C.M. (1991) Solution structure of neu- ronal bungarotoxin determined by two-dimensional NMR spec- troscopy: sequence specific assignments, secondary structure and dimer formation. Biochemistry 30, 4901–4909.

16. Peng, S.S., Kumar, T.K.S., Jayaraman, G. & Chang, C.C.C. (1997) Solution structure of toxin b, a long neurotoxin from the venom of the king cobra (Ophiophagus hannah). J. Biol. Chem. 272, 7817–7823.

Acknowledgements

17. Lin, S.R., Chi, S.H., Chang, L.S., Kuo, K.W. & Chang, C.C. (1996) Chemical modification of cationic residues in toxin a from King cobra (Ophiophagus hannah) venom. J. Protein Chem. 15, 95–101.

The authors wish to acknowledge and appreciate the financial support by Tehran University’s Research Council in the course of this research project. They also express their sincere thanks to Mr M. Erfani and the NMR facility at Tarbiat Modares University for their valuable assistance with NMR spectra. 18. Grishin, E.V., Sukhikh, A.P., Slobodyan, L.N., Ovchinnikov, Y.A. & Sorokin, V.M. (1974) Amino acid sequence of neurotoxin I from Naja naja oxiana venom. FEBS Lett. 45, 118–121.

References

19. Piotto, M., Saudek, V. & Sklenar, V. (1992) Gradient-tailored excitation for single-quantum NMR spectroscopy of aqueous solutions. J. Biomol. NMR 2, 661–665.

1. Dufton, M.J. & Hider, R.C. (1983) Conformational properties of the neurotoxins and cytotoxins isolated from Elapid snake venom. CRC. Crit. Rev. Biochem. 14, 113–171.

2. Endo, T. & Tamiya, N. (1987) Current view on structure function relationship of postsynaptic neurotoxins from snake venoms. Pharmacol. Ther. 34, 403–451. 20. Sklenar, V., Piotto, M., Leppik, R. & Saudek, V. (1993) Gradient- tailored water suppression for 1H-15N HSQC experiments opti- mized to retain full sensitivity. J. Magn. Reson. 102, 241–245. 21. Bax, A. & Davis, D.G. (1985) MLEV-17 based two-dimensional homonuclear magnetization transfer spectroscopy. J. Magn. Reson. 65, 355–360.

3. Joubert, F.J. (1973) Snake venom toxins, the amino acid sequences of two toxins from Ophiophagus hannah. Biochem. Biophys. Acta 317, 85–98.

22. Derome, A.E. & Williamson, M.P. (1990) Rapid-pulsing artifacts in double-quantum-filtered COSY. J. Magn. Reson. 88, 177–185. 23. Bartels, C., Xia, T.H., Billeter, M., Gu¨ ntert, P. & Wu¨ thrich, K. (1995) The program XEASY for computer-supported NMR spectral analysis of biological macromolecules. J. Biomol. NMR 5, 1–10.

4. Zinn-Justin, S., Roumestand, C., Gilquin, B., Bostem, F., Me´ nez, A. & Toma, F. (1992) Three-dimensional solution structure of curaremimetic toxin from Naja nigricillis venom: a proton NMR and molecular modeling study. Biochemistry 31, 11335–11347.

24. Szyperski, T., Guntert, P., Otting, G. & Wu¨ thrich, K. (1992) Determination of scalar coupling constants by inverse Fourier transformation of in-phase multiplets. J. Magn. Reson. 99, 552– 560. 5. Ruan, K.H., Stiles, B.G. & Atassi, M.Z. (1991) The short neuro- toxin binding regions on the a-chain of human and Torpedo cali- fornica acetylcholine receptors. Biochem. J. 274, 849–854.

6. Basus, V.J., Song, G. & Harrot, E. (1993) NMR solution structure of an a-bungarotoxin/nicotinic receptor peptide complex. Bio- chemistry 32, 12290–12298. 25. Wishart, D.S. & Sykes, D.W. (1994) The 13C chemical shift index: a simple method for the identification of protein secondary structure using 13C chemical-shift data. J. Biomol. NMR 4, 171– 180.

26. Cornilescu, G., Delaglio, F. & Bax, A. (1999) Protein backbone angle restraints from searching a database for chemical shift and sequence homology. J. Biomol. NMR 13, 289–302.

7. Ruan, K.H., Surlino, J., Quicho, F.A. & Atassi, A.M. (1990) Acetylcholine receptor–a-bungarotoxin interactions: determin- ation of the region-to-region by peptide–to–peptide interactions and molecular modeling of the receptor activity. Proc. Natl Acad. Sci. USA 87, 6156–6160.

8. Tsetlin, V.I. & Hucho, F. (2004) Snake and snail toxins acting on nicotinic acetylcholine receptors: fundamentals aspects and medical applications. FEBS Lett. 557, 9–13. 27. Bru¨ nger, A.T., Adams, P.D., Marius-Clore, G., DeLano, W.L., Gros, P., Grosse-Kunstleve, R.W., Jian-Sheng, J., Kuszewskic, J., Nilges, M., Navraj, S., Pannu, N.S., Randy, J., Read, R.J., Rice, L.M., Simonson, T. & Warren, G.L. (1998) Crystallography and NMR System: a New Software Suite for Macromolecular Struc- ture Determination. Acta Crystallogr. D. 54, 905–921.

28. Linge, J.P., O’Donoghue, S.I. & Nilges, M. (2001) Automated assignment of ambiguous nuclear overhauser effects with ARIA. Methods Enzymol. 339, 71–90. 9. Love, R.A. & Stroud, R.M. (1986) The crystal structure of a-bungarotoxin at 2.5 A˚ resolution: relation to solution structure and binding to acetylcholine receptor. Protein Eng. 1, 37–46. 10. Betzel, C., Lange, G., Pal, G.P., Wilson, K.S., Maelicke, A. & Saenger, W. (1991) The refined crystal structure of a-cobratoxin

NTX-1 solution structure from N. naja oxiana (Eur. J. Biochem. 271) 4957

(cid:1) FEBS 2004

39. Antil, S., Servent, D. & Me´ nez, A. (1999) Variability among the sites by which curaremimetic toxins bind to Torpedo acetylcholine receptor, as revealed by identification of the functional residues of a-cobratoxin. J. Biol. Chem. 274, 34851–34858. 29. Laskowski, R.A., Rullmann, J.A.C., MacArthur, M.W., Kaptein, R. & Thornton, J.M. (1996) AQUA and PROCHECK-NMR: Programs for checking the quality of protein structures solved by NMR. J. Biomol. NMR 8, 477–486.

30. Koradi, R., Billeter, M. & Wu¨ thrich, K. (1996) MOLMOL: a program for display and analysis of macromolecular structures. J. Mol. Graph. 14, 51–55.

40. Antil-Delbeke, S., Gaillard, C., Tamiya, T., Corringer, P.J., Changeux, J.P., Servent, D. & Me´ nez, A. (2000) Molecular de- terminants by which a long chain toxin from snake venom inter- acts with the neuronal a7-nicotinic acetylcholine receptor. J. Biol. Chem. 275, 29594–29601.

31. Case, D.A., Pearlman, D.A., Caldwell, J.W., Cheatham, T.E. III, Wang, J., Ross, W.S., Simmerling, C.L., Darden, T.A., Merz, K.M., Stanton, R.V., Cheng, A.L., Vincent, J.J., Crowley, M., Tsui, V., Gohlke, H., Radmer, R.J., Duan, Y., Pitera, J., Massova, I., Seibel, G.L., Singh, U.C., Weiner, P.K. & Kollman, P.A. (2002) AMBER 7. University of California, San Francisco. 41. Fruchart-Gaillard, C., Gilquin, B., Antil-Delbeke, S., Le Novere, N., Tamiya, T., Corringer, P.J., Changeux, P.J., Me´ nez, A. & Servent, D. (2002) Experimentally based model of a complex be- tween a snake toxin and the a7 nicotinic receptor. Proc. Natl Acad. Sci. USA 99, 3216–3221. 32. Wu¨ thrich, K. (1986) NMR of Proteins and Nucleic Acids. Wiley, New York.

42. Endo, T., Nakanishi, M., Furukawa, S., Joubert, F.J., Tamiya, N. & Hayashi, K. (1986) Stopped-flow fluorescence studies on bind- ing kinetics of neurotoxins with acetylcholine receptor. Biochem- istry 25, 395–404. 33. Joubert, F.G. (1973) The amino acid sequences of two toxin from Ophiophagus hannah (King cobra) venom. Biochem. Biophys. Acta 317, 85–98.

43. Searle, M.S. & Williams, D.H. (1992) The cost of conformational order: entropy changes in molecular associations. J. Am. Chem. Soc. 114, 10690–10697. 34. Mebs, D., Narita, K., Iwanaga, S., Samejika, Y. & Lee, C.Y. (1972) Purification, properties and amino acid sequences of a-bungarotoxin from the venom of Bungarus multicinctus. Hoppe- Seyler Z. Physiol. Chem. 353, 243–262.

44. Connolly, P.J., Stern, A.S., Turner, C.J. & Hoch, J.C. (2003) Molecular dynamics of the long neurotoxin LSIII. Biochemistry 42, 14443–14451. 35. Karlsoon, E., Eaker, D. & Ponterius, G. (1972) Modification of amino group in Naja naja neurotoxin and preparation of radio- active derivatives. Biochem. Biophys. Acta 257, 235–248.

36. Maeda, N. & Tamiya, N. (1974) Isolation, properties and amino acid sequences of the toxin Laticauda semifasciata III, a weak and reversibly acting neurotoxin from venom of a sea snake, Laticauda semifasciata. Biochem. J. 141, 389–400.

45. Bystrov, V.F., Tsetlin, V.I., Karlsson, E., Pashkov, V.S., Utkin, Y., Kondakov, V.I., Plushnikov, K.A., Arseniev, A.S., Ivanov, V.T. & Ovchinnikov, Y.A. (1983) Magnetic resonance evaluation of snake neurotoxin structure-function relationship. In Toxins as Tools in Neurochemistry (Hucho, F. & Ovchinnikov, Y.A., eds), pp. 193–233. Proceeding of the Symposium held at Berlin, 1983, Walter de Gruyter, Berlin.

Supplementary material

The following material is available from http://www. blackwellpublishing.com/products/journals/suppmat/EJB/ EJB4465/EJB4465sm.htm Table S1. The chemical shift table of NTX-1.

37. Scarselli, M., Spiga, O., Ciutti, A., Bernini, A., Bracci, L., Lelli, B., Lozzi, L., Calamandrei, D., Di Maro, D., Klein, S. & Niccolai, N. (2002) NMR structure of alpha-bungarotoxin free and bound to a mimotope of the nicotinic acetylcholine receptor. Biochemistry 41, 1457–1463.

38. Servent, D., Winckler-Dietrich, V., Hu, H.Y., Kessler, P., Drevet, P., Bertrand, D. & Me´ nez, A. (1997) Only snake curaremi- metic toxins with a fifth disulfide bond have high affinity for the neuronal alpha7 nicotinic receptor. J. Biol. Chem. 272, 24279– 24286.