Drysdalin, an antagonist of nicotinic acetylcholine receptors highlights the importance of functional rather than structural conservation of amino acid residues

Abstract Snake venom neurotoxins are potent antagonists of nicotinic acetylcholine receptors (nAChRs). Here, we describe a novel member of class 3c long‐chain neurotoxin drysdalin from the venom of Drysdalia coronoides. Drysdalin lacks three of the eight conserved classical functional residues critical for nAChRs interaction. Despite such a drastic alteration of the functional site, recombinant drysdalin showed irreversible postsynaptic neurotoxicity with nanomolar potency and selectively antagonizes the rodent muscle (α1)2β1δε, and human α7 and α9α10 nAChRs, but had no significant activity at the human α3β2, α3β4, α4β2, and α4β4 nAChRs. Substitution of Leu34 and Ala37 residues with the conserved Arg had minimal impact on the potency whereas conserved Phe replacement of residue Arg30 substantially reduced or abolished inhibitory activity. In contrast, truncation of the 24‐residue long C‐terminal tail leads to complete loss in (a) activity at α9α10 nAChR; and (b) irreversibility with reduced potency at the muscle and α7 nAChRs. Overall, the non‐conserved Arg30 residue together with the uniquely long C‐terminal tail contribute to the inhibitory activity of drysdalin at the nAChRs suggesting, at least for drysdalin, functional rather than sequence conservation plays a critical role in determining the activity of the toxin.

Many nAChR agonists and antagonists are isolated from plants (nicotine, d-tubocuraine, methyllycaconitine), bacteria (invermectin), algae (anatoxin-a), and animals such as cone snails (conotoxins), corals (lophotoxin), toads (epibatidine), and snakes (α-bungarotoxin [Bgtx]) (reviewed in 7 ). Snake venom is one of the most abundant sources of natural ligands interacting with nAChRs with high specificity and potency. One such toxin isolated from Bungarus multicinctus is Bgtx. As with Bgtx, most snake venom-derived nAChRs antagonists belong to the three-finger toxin (3FTx) family with three β-stranded "fingers" extending from a central core containing four conserved disulphide bridges. Although members of this family share a common fold, they bind to distinct receptors, enzymes, and ion channels, and exhibit a wide variety of biological effects utilizing unique functional sites (reviewed in 8 ). Neurotoxins are one of the largest members of the 3FTx family that interfere with cholinergic transmission in the central and peripheral nervous systems.
The commonly accepted dogma that all LNTXs with a fifth disulphide block neuronal α7 nAChRs was challenged by some "classical" long-chain α-neurotoxins that have poor binding affinity for α7 nAChR. 19,20 The authors proposed that Arg33Leu and Phe29Arg substitutions in the "neuronal pharmacophore" region may be responsible for the lack of affinity. Thus, the substitutions in the key functional residues may lead to almost complete loss of binding affinity.
In our previous study, the venom of an Australian elapid snake Drysdalia coronoides was profiled to be rich in 3FTx and serine protease inhibitors. 21 A variety of 3FTxs had sequences homologous to SNTX-and LNTX-postsynaptic α-neurotoxins, with several conserved functional residues involved in recognizing nAChRs. Among them, five 3FTx isoforms, represented by LNTX 13,43,173,346,and 146R were the longest (mature protein of 87 aa residues) of all the 3FTxs. This particular group of 3FTxs have unusually long C-terminal tail (24 aa residues). Furthermore, three functionally important residues were substituted by non-conserved residues, namely Phe30Arg, Arg34Leu, and Arg37Ala. In the first substitution, an aromatic, hydrophobic residue is replaced by a charged, hydrophilic residue, whereas in the latter two substitutions, a positively charged, hydrophilic residue is replaced by an aliphatic, hydrophobic residue. Thus, they belong to a distinct class of LNTXs. We chose to work on the most abundant isoform, LNTX 13. Here we report the recombinant expression and purification, and pharmacological and electrophysiological characterization of drysdalin (Drysdalia toxin) -the first member of this novel group of 3FTxs. Interestingly, despite the substitution of functionally conserved residues in the loop II, drysdalin exhibits nanomolar postsynaptic irreversible neurotoxicity. It inhibits adult rodent muscle α1β1δε and human neuronal α7 and α9α10 nAChRs with nanomolar potency. Furthermore, we describe the roles of the C-terminal tail and the three non-conserved residues of drysdalin in its interaction with muscle and neuronal nAChRs.

| Recombinant protein expression, purification, and refolding
Drysdalin and its mutants were cloned into a pET-32a modified vector and expressed as inclusion bodies (IBs) in Escherichia coli SHuffle ® cells. The IBs were solubilized in denaturing | 117 CHANDNA et Al.

| Ex vivo CBCM organ bath
All experiments were conducted according to the Protocol (103/08A) approved by the National University of Singapore Institutional Animal Care and Use Committee. Isolated chick biventer cervicis muscle (CBCM) experiments were conducted in a 6 ml-organ bath chamber, containing Krebs-Henseleit buffer bubbled with carbogen (5% CO 2 in O 2 ) at 37°C, as described previously. 22 Neuromuscular blockade by drysdalin is expressed as percentage of the twitch height in the absence of drysdalin to the twitch height 30 minutes post exposure to drysdalin. The half-maximal inhibitory concentration (IC 50 ) was determined from concentration-response curve fitted to a nonlinear regression function and reported with error of the fit. To test for reversibility, recovery from complete neuromuscular blockade by drysdalin was assessed by washing the tissue with Krebs solution over a 120-minute period after 80% blockade of the twitch responses.

| Electrophysiology
RNA preparation, oocyte preparation, and expression of nAChRs were performed as described previously. 23 All protocols were approved by the University of Sydney Animal Ethics Committee. Briefly, stage V-VI oocytes were obtained from Xenopus laevis, and defolliculated with 1.5 mg/mL Type II collagenase in OR-2 solution. To conduct the twoelectrode voltage clamp (TEVC) experiments on the oocytes, nAChRs were expressed by microinjecting the cRNA of various receptor subunits into oocytes. Electrophysiological recordings were carried out 2-7 days after microinjection. Oocytes expressing nAChRs were voltage clamped at a holding potential −80 ,mV. 24 nAChR-mediated currents were evoked by applying ACh for 1.5 seconds at a rate of

Class 2b
P01380|α-EPTX-Ast2a 2 mL/min, at a half-maximal effective concentration (EC 50 ) of each subtype followed by washouts of 180 seconds between ACh applications. Oocytes were incubated with the toxin for 5 minutes before ACh was co-applied. Peak AChevoked current amplitude before and after toxin incubation was recorded using pClamp 9 software and measured using Clampfit 10.7 (Molecular Devices, Sunnyvale, CA, USA). Concentration-response curves were used to determine the IC 50 of the toxin at each receptor.

| Classification of LNTXs
Curare-mimetic α-neurotoxins have been classified by Endo and Tamiya into SNTXs and LNTXs. 12 Both SNTXs and LNTXs share a similar three-dimensional structure of 3FTxs that is held together by four conserved disulphide bonds.
LNTXs have an additional fifth disulphide bond in their second loop. Although both SNTXs and LNTXs bind to the muscle nAChR subtype (picomolar), only LNTXs inhibit neuronal α7 nAChRs with high affinity (nanomolar).
Using systematic site-directed mutagenesis of 29 residues, Antil et al explored the role of all Cbtx three loops and identified common residues (Trp25, Asp27, Phe29, Arg33, Arg36, and Phe65) that interact with both Torpedo and α7 nAChRs and receptor subtype-selective residues (Lys23 and Lys49 at Torpedo nAChR and Ala28, Lys35 and the Cys26-Cys30 disulphide bond at α7 nAChR). 15,16 In the last two decades, a significant number of LNTXs have been discovered with minimal characterization of residues at the functional site. Our preliminary evaluation indicated that some of these LNTXs have distinct structural differences and therefore, we classified them further into three classes based on the number of C-terminal tail residues after the last Cys residue in the aa sequence and variations in the functionally important residues (Figures 1 & 2).

| Class 1
Toxins with the shortest chain length (66-68 aa residues) and only 4-5 residues in the C-terminus were classified into Class 1 ( Figure 1A). They were further categorized as 1a and 1b based on the differences in the functionally important residues in loop II, III, and C-terminus. We observed that two functionally important residues were substituted by non-conserved residues, namely, Lys53 to Glu and Phe70 to His. Lys53Glu substitution reduced Cbtx affinity for both binding sites at the Torpedo nAChR, with one site greatly reduced by 52-fold whereas the other by only threefold. 16 This substitution also led to threefold reduction in affinity for the Gallus α7/5-HT 3 chimeric receptor. 15 The impact of Phe70His substitution on the affinity is unclear as Cbtx and Bgtx have Phe and His at the homologous position, respectively. The functional impact of augmenting the short C-terminus has been studied in both Cbtx and Bgtx. A threefold reduction in the RKCYKTHP---YKSEPCAPGENLCYTKTWCDFRCSQLGKAVELGCAATCPTTKP-YEEVTCCSTDDCNRFPNWERPRPRPRGLLSSIMDHP affinity for the α7/5-HT 3 nAChR chimera was reported for the Cbtx C-terminal deletion mutant Cbtx[ΔPro66-Pro71] 15 whereas, truncation of Bgtx seven C-terminal residues (Bgtx[ΔHis68-Gly74]) decreased the apparent binding affinity to the Torpedo nAChR by sevenfold. 18

| Class 2
Toxins with intermediate chain length (68-75 aa residues) and 6-13 C-terminal residues were classified into Class 2 ( Figure 1B). They were further categorized into four subclasses (2a-d) based on the differences in the functionally important residues in loop II, III, and C-terminus. Class 2a toxins share a common His70 residue with 6-9 residues in the C-terminus, whereas Class 2c toxins have a conserved Phe residue at position 70 with slightly longer C-terminal tail consisting of 7-13 residues (Figure 2). Class 2b toxins have either Tyr/Val/Pro residue at position 70. Class 2d toxins share common Arg33 and Leu37 residues, whereas toxins in subclasses 2a-c have an aromatic residue and a positively charged Arg residue at position 33, 37, respectively. The key feature of most toxins in groups 2a and 2c is the presence of two positively charged Lys and Arg residues at the C-terminus. Removal of the corresponding Lys70 and Arg72 residues in Bgtx[ΔHis68-Gly74] resulted in a sevenfold reduction in the binding activity at the Torpedo nAChR. 18

| Class 3
Toxins with the longest chain length (79-88 aa residues) and the longest C-terminus (17-24 aa residues) were classified into Class 3 ( Figure 1C). They were further divided into three subclasses (3a-c) based on the length and sequence identity of the C-terminus. Class 3a toxins have the shortest tail among Class 3 toxins with only 17 aa residues in the C-terminal tail. Classes 3b and 3c share similar C-terminal sequences, however, LNTXs of class 3c have additional 4-6 residues (mostly Pro and Arg residues) upstream of the homologous C-terminal region. For majority of class 3c toxins, key functionally important residues Phe/Trp33 and Arg37, and Arg40 are substituted with Arg/Leu and Ala/Leu, respectively ( Figure 2). Drysdalin is grouped in class 3c α-neurotoxin (labelled DrycnLNTX-13 in Figure  1C) with a long 24-aa C-terminal tail and the consensus LNTX residues Phe30, Arg34, and Arg37 are substituted by residues Arg, Leu and Ala, respectively. The importance of these residues positions has been explained above. Interestingly, the length of class 3c toxins are similar to those of nAChR interacting human Ly6 proteins (Lynx1, PATE4, SLURP1, and SLURP2), 25 but they share poor sequence similarity (19%-25%) and class 3c toxins have much longer C-terminal tail segments.
To date, the function of only Cbtx (Class 2c) and Bgtx (Class 2a) has been well characterized. The impact of several key functional residue substitutions and changes in the C-terminal segments on the nAChR subtype selectivity of class 3c members are unknown. Here, we characterize the first member of class 3c α-neurotoxin, drysdalin (DrycnLNTX-13). In this toxin, three out of the eight functional residues (37.5%) are replaced by non-conserved residues, and additionally it has a 24 aa residue-long C-terminal tail. Here, we describe the function of drysdalin and the role of these non-conserved residues and C-terminal tail in the interactions with various nAChRs.

| Recombinant expression and purification of drysdalin
The venom yield of D. coronoides is typically 2-3 mg per milking (Mr Peter Mirtschin, Venom Supplies Pty Ltd., Tanunda, South Australia; unpublished observations). Due to the low venom yield, we recombinantly expressed drysdalin to obtain large quantities of the protein. Drysdalin was expressed predominantly as IBs ( Figure 3A). The Triton X-100washed IBs were solubilized in a denaturing buffer, reduced with DTT, and purified to homogeneity by reverse phase-high performance liquid chromatography (RP-HPLC) ( Figure  3B). ESI-MS data of reduced drysdalin showed six peaks of mass/charge (m/z) ratios ranging from +6 to +11 ( Figure  3C). The reconstructed mass spectrum showed a molecular mass of 11731.63 ± 0.56 Da (Figure 3C inset), matching the calculated 11732.2 Da mass. After refolding, all the conformers were separated on a Jupiter semi-preparative C18 column ( Figure 3D) and the major peak was rechromatographed on an analytical C18 column to homogeneity (Figure 3D inset). ESI-MS data of refolded drysdalin showed six peaks of mass/ charge (m/z) ratios ranging from +6 to +11 ( Figure 3E). The reconstructed mass spectrum of the refolded drysdalin showed a molecular mass of 11721.87 ± 0.63 Da ( Figure  3E inset), indicating a loss of ten protons due to the formation of five disulphide bonds. The CD spectrum of drysdalin showed a minimum at 206-209 nm, indicating the presence of β-sheeted secondary structure characteristic of 3FTxs ( Figure  3F).

| Drysdalin is a potent postsynaptic neurotoxin
To examine the biological effects of drysdalin, the refolded recombinant drysdalin was injected intraperitoneally into adult mice at a dose of 0.25 and 3 mg/kg of body weight. Both mice showed typical signs of peripheral neurotoxicity such as hind limb paralysis, laboured breathing, and finally death (after 46 minutes at 3 mg/kg) presumably due to respiratory paralysis. 26,27 No gross changes were observed in the internal organs of the mice.

| Drysdalin inhibits heterologously expressed rodent and human nAChRs
As LNTXs bind to both muscle and neuronal α7 nAChRs, 12 we tested the activity of drysdalin on nAChRs heterologously expressed in X. laevis oocytes. Drysdalin selectively blocked the rodent (r) muscle-type α1β1δε and human (h) neuronal α7 and α9α10 subtypes, and had no significant inhibitory activity at hα3β2, hα3β4, hα4β2, and hα4β4 nAChRs when tested up to 30 nmol/L ( Figure 5).

F I G U R E 6
Drysdalin inhibition of rα1β1δε, hα7, and hα9α10 nAChRs subtypes. A, Representative ACh (1 μmol/L)-evoked currents recorded from Xenopus oocytes expressing rodent muscle α1β1δε nAChRs in the absence (control) and presence of 1-300 nmol/L drysdalin (top row), and following washout between applications of drysdalin (bottom row). B, Concentration-response curve of relative ACh-evoked current amplitude mediated by rα1β1δε nAChRs in the presence of drysdalin gave an IC 50 of 16.9 ± 3.6 nmol/L (n = 3) and Hill coefficient (n H ) of 1.3. C, Representative ACh (200 μmol/L)-evoked currents recorded from oocytes expressing human (h)α7 nAChR in the absence (control) and presence of 1-300 nmol/L drysdalin (top row), and following washout between applications of drysdalin (bottom row). D, Concentration-response curve of relative ACh-evoked current amplitude mediated by hα7 nAChR in the presence of drysdalin gave an IC 50 of 9.98 ± 0.6 nmol/L (n = 5) and n H of 1.9. E, Representative ACh (6.5 μmol/L)-evoked currents recorded from oocytes expressing hα9α10 nAChRs in the absence (control) and presence of 1-300 nmol/L drysdalin (top row), and following washout between applications of drysdalin (bottom row). F, Concentration-response curve of relative ACh-evoked current amplitude mediated by hα9α10 nAChRs in the presence of drysdalin gave an IC 50 of 11.3 ± 0.2 nmol/L (n = 6-9) and n H of 1.1 with a Hill coefficient of 1.3 indicating similar affinities to the two ACh binding sites (α1/δ, α1/ε) 29 (Figure 6A-B). In comparison to drysdalin, the two other LNTXs, Bgtx and Cbtx, also irreversibly inhibited the rα1β1δε nAChRs with similar (IC 50 = 14.0 ± 1.7 nmol/L) and 13-fold higher potency (1.3 ± 0.1 nmol/L), respectively (Table 1).
At hα9α10 nAChR, drysdalin concentration-dependently and reversibly inhibited ACh-evoked currents with an IC 50 of 11.3 ± 0.2 nmol/L ( Figure 6F). Bgtx and Cbtx also inhibited hα9α10 nAChRs in a concentrationdependent manner with an IC 50 of 9.7 ± 0.7 nmol/L and 7.9 ± 0.7 nmol/L, respectively (Table 1), and most importantly, the inhibition of hα9α10 nAChR was reversible for both LNTXs. The reversibility of drysdalin inhibition of hα9α10 contrasts with the non-reversible inhibition at rα1β1δε and hα7 nAChRs which may arise from differences in the binding site residues at the subunit interfaces (Supporting Information Figure S5).

| C-terminal tail plays a critical role in the activity of drysdalin
Residues in the C-terminus of Bgtx and Cbtx have been implicated in interactions with the nAChRs. Both Bgtx[ΔHis68-Gly74] 18 and Cbtx[ΔPro66-Pro71] 15 C-terminus deleted analogues had reduced apparent affinity at the muscle-type Torpedo nAChR (sevenfold) and Gallus α7/5-HT 3 chimeric receptor (threefold), respectively. Additionally, alanine substitution of the Cbtx C-terminal residues resulted in reduced affinities at the aforementioned receptors. At the Torpedo nAChR, Cbtx[Phe65Ala] had sevenfold affinity loss 16 whereas at the α7/5-HT 3 chimera, the affinities of both Cbtx[Phe65Ala] and Cbtx[Pro66Ala] were reduced by 16-and threefold, respectively. 15 Drysdalin has an extended 24 residue-long C-terminal tail compared to other LNTXs ( Figure 1C). To determine the contribution of the C-terminal tail in interactions with nAChRs, truncated drysdalin (tDrysdalin) lacking 20 aa residues of the C-terminus was designed and recombinantly expressed. (Supporting Information Figures S2, S3D). The refolded truncated protein showed mass corresponding to the calculated value (Supporting Information Table S1) and the overall structure of tDrysdalin was similar to the full-length protein (Supporting Information Figure S4).
The effect of Bgtx and Cbtx truncation on the reversibility of ACh current inhibition is currently unknown. Here, we report for the first time the effect of C-terminal residue removal on the reversibility of any three-finger α-neurotoxin on the inhibition of ACh-evoked currents mediated by nAChRs.

| "Reverting" mutations of nonconserved functional residues has no significant effect on the activity of drysdalin
Structural studies of LNTXs strongly suggest the involvement of three conserved residues (one Phe and two Arg) in binding of the toxins to their molecular targets. The interaction of these residues was observed in the crystal structure of Lymnaea stagnalis (Ls)-acetylcholine binding protein (AChBP) with Cbtx. Ls-AChBP principal subunit residues Tyr185 and Tyr192 form π-π interactions with Cbtx Phe29 and additionally Tyr185 interacts with Cbtx Arg33 (1st Arg) and Arg36 (2nd Arg) via cation-π interactions (Supporting Information Figure S6A). 30 Similar cation-π interactions are also energetically favourable in T A B L E 1 IC 50 values of drysdalin, α-bungarotoxin (Bgtx) and α-cobratoxin (Cbtx) at rodent (r)α1β1δε, human (h)α7, and hα9α10 nAChRs  the binding of Cbtx to an 18-mer cognate peptide derived from the Torpedo muscle nAChR (Supporting Information Figure S6B). 31 An elaborate cation-π interaction system is also involved in the interaction between Bgtx loop II residues Phe32 and Arg36 (1st Arg) with residues Trp93, Tyr190, and Tyr198 of the monomeric mouse α1 nAChR extracellular domain (Supporting Information Figure S6C). 32 At the α7/Ls-AChBP chimera, Bgtx loop-II residues Arg36 and Phe32 form a πcation stack that aligns edge-to-face with the conserved α7-loop C Tyr184 and additionally, Arg36 interacts with receptor residues Tyr91, Trp145 and Tyr191 (Supporting Information Figure S6D). 33 Mutational studies show that Tyr184 alone is required for the high affinity of Bgtx to α7 receptor as seen from the complete loss in the affinity for the mutation Tyr184Thr. Single residue mutations Tyr184Phe and Tyr191Thr do not affect Bgtx apparent affinity at the receptor. However, with both mutations the apparent affinity decreases by sixfold. 34 Hence, conserved residues Phe and Arg of Cbtx and Bgtx form an extensive cation-π interaction system in the aromatic ligand-binding pocket of muscle as well as neuronal nAChRs. However, there are subtle  Table 3 differences in the interactions of LNTXs with different subtypes of nAChRs.
Structure-function relationship studies of LNTXs also indicate the importance of Phe and Arg residues in recognizing the muscle and neuronal nAChRs. Cbtx loop II mutants [Phe29Leu/Ala], [Arg33Glu], and [Arg36Ala] had decreased binding to the Torpedo nAChR and α7/5-HT 3 receptor. 15,16 In contrast, substitution of Phe29 with conserved Trp residue did not affect the affinity to both Torpedo and neuronal receptors (1.29-and 0.7-fold, respectively).
The importance of the conserved LNTX Phe and Arg residues in high affinity binding to the neuronal α7 nAChRs is further substantiated by the weak inhibitory activities of α-elapitoxin-Aa2a of Acanthophis antarcticus (Phe and 1st Arg are substituted by Arg29 and Leu33, respectively) 19 and α-elapitoxin-Al2a of Austrelaps labialis (only 2nd Arg is replaced by Val36). 20 Intriguingly, drysdalin lacks the conserved LNTX Phe (replaced by Arg at position 30) and Arg residues (replaced by Leu and Ala at position 34 and 37, respectively), which is three out of the eight functionally conserved residues (see Section 1; Figure 1C). Despite these significant disparities, drysdalin exhibits nanomolar potency to muscle as well as neuronal nAChRs (Table 1). We speculated that reverting mutation of these non-conserved residues to the functionally conserved residues would enhance the antagonistic activity of drysdalin at the muscle and neuronal nAChRs.
Single  Figure S1) were generated by one-step site directed plasmid mutagenesis protocol, 35 and recombinantly expressed in E coli, folded and purified by RP-HPLC (Supporting Information Figures S2, S3). The reduced and refolded proteins showed the mass corresponding to the calculated values (Supporting Information Table S1). The CD spectra of the folded mutants were similar to the full-length drysdalin (Supporting Information Figure S4), indicating that the mutations did not affect the overall secondary structure of the toxin and presumably the three-finger fold.

| Drysdalin potency at inhibiting AChevoked currents of nAChRs is dependent on non-conserved residues
At the muscle nAChR, Drys[R30F] was ~fourfold less potent (IC 50 = 74.6 ± 6.7 nmol/L) than drysdalin (IC 50 = 16.9 ± 3.6 nmol/L) ( Figure 8A, Table 3). Aromatic residues (Phe/Trp) at this position are conserved in all LNTXs except for classes 2d, 3a and 3c ( Figure 1) and play a crucial role in interacting with Tyr185, Tyr192 of the muscle nAChR through π-π interactions. In class 2d, 3a and 3c of LNTXs (except 476539402), the conserved Phe/Trp is replaced by Arg resulting in cation-π interactions between the aromatic nAChR residues and the guanidinium group of Arg. 36 The reduced potency of Drys[R30F] suggests that cation-π interaction is preferred over π-π interaction in the ligand-binding pocket of the muscle nAChR.

| Implication on homology and understanding of protein chemistry
Multiple sequence alignment is one of the widely used bioinformatics tool used to determine structure-function relationships, structural modelling, functional prediction, phylogenetic analysis, and sequence database searching. Structural and functional similarities are predicted based on the identity and similarity between the residues of the proteins where the higher the identity/similarity, the higher the confidence in predictions and conversely, nonconserved substitutions lead to loss of structure or function. Here we describe for the first time, for drysdalin non-conserved substitutions continue to retain the functional efficiency of the protein as they maintain similar interactions through altered modes/mechanisms. Our results indicate that further considerations are needed to enhance the functional predictability of proteins using bioinformatics approaches.

| CONCLUSIONS
In LNTXs, eight conserved residues define the binding of the toxins to nAChRs and confer potent postsynaptic neurotoxicity. 15,16 Uniquely, the conserved one Phe and two Arg residues are absent in drysdalin, replaced by Arg30 and, Leu34 and Ala37, respectively. It selectively antagonizes rodent muscle (α1) 2 β1δε, and human α7 and α9α10 nAChRs. Replacing Leu34 and Ala37 residues with the conserved LNTX Arg residue had minimal impact on the potency at the nAChRs. In contrast, substituting Arg30 residue with Phe impaired the inhibitory activity of drysdalin. In addition to the residues in loop II, we found that the C-terminal tail plays a critical role in recognizing α9α10 nAChR. The C-terminal tail also affects the reversibility of drysdalin at the muscle and α7 nAChRs. Thus, although the conservation of functional residues in a protein family is important in retaining its function, it may not be absolute.