Residues Responsible for the Selectivity of α-Conotoxins for Ac-AChBP or nAChRs

Nicotinic acetylcholine receptors (nAChRs) are targets for developing new drugs to treat severe pain, nicotine addiction, Alzheimer disease, epilepsy, etc. α-Conotoxins are biologically and chemically diverse. With 12–19 residues and two disulfides, they can be specifically selected for different nAChRs. Acetylcholine-binding proteins from Aplysia californica (Ac-AChBP) are homologous to the ligand-binding domains of nAChRs and pharmacologically similar. X-ray structures of the α-conotoxin in complex with Ac-AChBP in addition to computer modeling have helped to determine the binding site of the important residues of α-conotoxin and its affinity for nAChR subtypes. Here, we present the various α-conotoxin residues that are selective for Ac-AChBP or nAChRs by comparing the structures of α-conotoxins in complex with Ac-AChBP and by modeling α-conotoxins in complex with nAChRs. The knowledge of these binding sites will assist in the discovery and design of more potent and selective α-conotoxins as drug leads.

From the crystal structure of five α-conotoxins in complex with Ac-AChBP, it was found that the five α-conotoxins bound to a similar position of the nearby loop C of the Ac-AChBP, and the binding pattern is also similar. The structures of α-conotoxins are largely determined by the conserved disulfide connectivity between CysI-CysIII and CysII-CysIV [47,49]. The mutations of α-conotoxin, which were observed in the Ala substitutions (no change of the disulfide connectivity), showed no major changes in secondary structure (except Pro residue) [19,38]; the α-conotoxin mutants have a well-defined helix that is similar to native conotoxins. From the structure of α-conotoxins in complex with Ac-AChBP, and as shown by many experiments, it is assumed that the α-conotoxin mutants were binding at a similar site of the receptor and displaying binding patterns similar to native conotoxins [47,49]. Most conotoxins share a common folding structure that is not expected to change upon binding, therefore "rigid" docking can be performed to determine the residues responsible for α-conotoxin's binding and selectivity for nAChRs [41,[50][51][52].
The structure of conotoxins bound by nAChRs is similar to the crystal structure of the conotoxins in complex with Ac-AChBP [36]. For instance, the backbone tilt of conotoxin PnIA (A10L D14K) bound by hα7 nAChR is similar to the crystal structure of PnIA (A10L D14K) in complex with Ac-AChBP (PDB: 2BR8) [36]. This is favorable for using computer models to design and discover new potent α-conotoxins [2,3,[51][52][53].
Here, we review the data of the co-crystal structure of α-conotoxins in complex with Ac-AChBP and some of the computer models of α-conotoxins selective for nAChRs, which will help to discover and design α-conotoxin analogs with higher selectivity for nAChRs. Asn-11 and Asn-12 are conserved in many α-conotoxins (shown in Figure 1A) [2]. From the crystal structure of α-conotoxins in complex with Ac-AChBP, it was found that Asn-11 and Asn-12 mainly bind to loop C of Ac-AChBP. A comparison of the Asn-11 and Asn-12 residues of the α-conotoxins in the crystal structure is shown in Figure 2. In Figure 2A, Asn-11 and Asn-12 of PnIA (A10L D14K), TxIA (A10L) and GIC form a hydrogen bond with the principal side ("+"-side, indicated in green, the residue-binding ligand is near loop C) [36] residues, Glu-191 and Tyr-91, and with the complementary side ("−"-side, indicated in blue, the residue-binding ligand is near loop E) [36,51] residues Arg-77. Mutating the residues of Asn-11 and Asn-12 to Ala resulted in the loss of affinity for Ac-AChBP [47]. Arg-11 of ImI is also a key residue for binding to Ac-AChBP. As shown in Figure 2B, the Arg-11 position of ImI is similar to the Asn-11 of GIC, which contacts the Ac-AChBP principal residue Glu-191 [45]. similar to native conotoxins [47,49]. Most conotoxins share a common folding structure that is not expected to change upon binding, therefore "rigid" docking can be performed to determine the residues responsible for α-conotoxin's binding and selectivity for nAChRs [41,[50][51][52].
Here, we review the data of the co-crystal structure of α-conotoxins in complex with Ac-AChBP and some of the computer models of α-conotoxins selective for nAChRs, which will help to discover and design α-conotoxin analogs with higher selectivity for nAChRs.

Asn-11, Asn-12 Residues of PnIA (A10L D14K), TxIA (A10L), and GIC Are the Key Residues for Binding to Ac-AChBP
Asn-11 and Asn-12 are conserved in many α-conotoxins (shown in Figure 1A) [2]. From the crystal structure of α-conotoxins in complex with Ac-AChBP, it was found that Asn-11 and Asn-12 mainly bind to loop C of Ac-AChBP. A comparison of the Asn-11 and Asn-12 residues of the α-conotoxins in the crystal structure is shown in Figure 2. In Figure 2A, Asn-11 and Asn-12 of PnIA (A10L D14K), TxIA (A10L) and GIC form a hydrogen bond with the principal side ("+"-side, indicated in green, the residue-binding ligand is near loop C) [36] residues, Glu-191 and Tyr-91, and with the complementary side ("−"-side, indicated in blue, the residue-binding ligand is near loop E) [36,51] residues Arg-77. Mutating the residues of Asn-11 and Asn-12 to Ala resulted in the loss of affinity for Ac-AChBP [47]. Arg-11 of ImI is also a key residue for binding to Ac-AChBP. As shown in Figure 2B, the Arg-11 position of ImI is similar to the Asn-11 of GIC, which contacts the Ac-AChBP principal residue Glu-191 [45].  [45] is similar to Asn-11 of GIC, which also contacts the Ac-AChBP principal side residue  The function of the Asn-11 and Asn-12 residues was also demonstrated by the α-conotoxin PnIA (A10L) truncation assay [54]. PnIA (A10L) loop 2 truncation indicated that removal of the residues (Pro-13, Asp-14, Tyr-15), after Asn-12 did not much affect PnIA (A10L) binding to rα7, but The function of the Asn-11 and Asn-12 residues was also demonstrated by the α-conotoxin PnIA (A10L) truncation assay [54]. PnIA (A10L) loop 2 truncation indicated that removal of the residues (Pro-13, Asp-14, Tyr-15), after Asn-12 did not much affect PnIA (A10L) binding to rα7, but when truncated to Asn-12, the PnIA (A10L) lost the activity to rα7 and acetylcholine-binding protein from Lymnaea stagnalis (Ls-AChBP). The sequence length of α-conotoxin ImI is similar to that of PnIA (A10L) as the removal of the residues (Asn-12, Pro-13, Asp-14, Tyr-15), but it has higher affinity to rα7 nAChR. The reason is that the residues Asp-5 and Arg-11 of ImI and the residues Asn-11 and Asn-12 of PnIA (A10L) have similar positions, respectively. Within the binding pocket of rα7, they can form hydrogen bonds with rα7 residues Glu-191, Tyr-193, and Tyr-91. So, ImI activity is actually similar to the full-length PnIA (A10L) [54]. His-5 in α-conotoxin GIC acts as a key residue for contacting with the hα3β2 nAChRs [47]; Arg-5 acts as a key residue, binding with rα3β2 nAChR [24]; and Arg-7 is a key residue of α-conotoxin ImI for its high affinity to hα7 nAChR [45]. Comparison of the crystal structures is shown in Figure 3. when truncated to Asn-12, the PnIA (A10L) lost the activity to rα7 and acetylcholine-binding protein from Lymnaea stagnalis (Ls-AChBP). The sequence length of α-conotoxin ImI is similar to that of PnIA (A10L) as the removal of the residues (Asn-12, Pro-13, Asp-14, Tyr-15), but it has higher affinity to rα7 nAChR. The reason is that the residues Asp-5 and Arg-11 of ImI and the residues Asn-11 and Asn-12 of PnIA (A10L) have similar positions, respectively. Within the binding pocket of rα7, they can form hydrogen bonds with rα7 residues Glu-191, Tyr-193, and Tyr-91. So, ImI activity is actually similar to the full-length PnIA (A10L) [54].

His-5 of GIC, Arg-5 of TxIA (A10L), and Arg-7 of ImI Are the Important Residues Responsible for α-Conotoxin's Selectivity for the Ac-AChBP Principal Side
His-5 in α-conotoxin GIC acts as a key residue for contacting with the hα3β2 nAChRs [47]; Arg-5 acts as a key residue, binding with rα3β2 nAChR [24]; and Arg-7 is a key residue of α-conotoxin ImI for its high affinity to hα7 nAChR [45]. Comparison of the crystal structures is shown in Figure 3. His-5 of GIC, Arg-5 of TxIA (A10L), and Arg-7 of ImI reside in similar positions. His-5 of GIC contacts Tyr-91 and Tyr-186; Arg-5 of TxIA (A10L) protrudes into the principal binding site and contacts Asp-195 and Tyr-186; and Arg-7 of ImI is also located in a similar position and contacts Tyr-91, Trp-145, IIe-194, and Asp-195. Mutation of Arg-7 to Ala in ImI resulted in a loss of affinity for Ac-AChBP [24,45].

Ser-4 of GIC, PnIA (A10L D14K), and ImI Are Key Residues for Interaction with the Ac-AChBP Complementary Side
Ser-4 is highly conserved in many α-conotoxins (shown in Figure 1A), which usually binds to the complementary side of Ac-AChBP. Comparison of Ser-4 in the crystal structure of different α-conotoxins is shown in Figure 4. In Figure 4A, it is shown that Ser-4 of GIC and PnIA (A10L D14K) are key residues for binding to the Ac-AChBP complementary side residues Asp-162, Ser-164, and Ser-165. Ser-4 of α-conotoxin TxIA (A10L), with a 20° backbone tilt rotation, makes only weak contact with Ser-165 of the Ac-AChBP complementary side [24,36,47]. Figure 4B shows Ser-4 of ImI located on the right side of Ser-4 of GIC and forms a hydrogen bond with Asp-162. Ser-4 of BulA is located on the left side of Ser-4 of GIC and forms a hydrogen bond with Ser-165 of the Ac-AChBP complementary side [36,47].  Ser-4 is highly conserved in many α-conotoxins (shown in Figure 1A), which usually binds to the complementary side of Ac-AChBP. Comparison of Ser-4 in the crystal structure of different α-conotoxins is shown in Figure 4. In Figure 4A, it is shown that Ser-4 of GIC and PnIA (A10L D14K) are key residues for binding to the Ac-AChBP complementary side residues Asp-162, Ser-164, and Ser-165. Ser-4 of α-conotoxin TxIA (A10L), with a 20 • backbone tilt rotation, makes only weak contact with Ser-165 of the Ac-AChBP complementary side [24,36,47].

α-Conotoxin Residues Selective for nAChRs
From the α-conotoxins complex with Ac-AChBP structure, and as shown in many experiments, it is assumed that α-conotoxin binding sites of nAChRs are similar to the binding sites of Ac-AChBP [1,36,49]. Sequence alignment of Ac-AChBP, Ls-AChBP, h(r)α3, h(r)α4, h(r)α6, h(r)α7, and h(r)α10 residues forming binding sites on the principal side are shown in Figure 6A. The residues forming binding sites on the complementary side are shown in Figure 6B. In Figure 6, it is shown that different nAChRs have different residues in the binding site, which could affect the conotoxin's affinity. For example, the Ile-57 (number 59 in that study [31]) of hα9 was replaced by Thr-57 in rα9 ( Figure 6B) result in α-conotoxin Vc1.1 increased affinity 10 folds for the rat receptor [31]. Modelling conotoxins binding to nAChRs could determine the key residues of α-conotoxins for a higher affinity to nAChRs [18,31,49]. Tyr-10 of ImI is also located in the pocket consisting of Arg-77, Val-106, Thr-108, Ser-112, and Met-114. Importantly, the substitution of Val-106 by Arg-106 in Ls-AChBP induces a steric clash with Tyr-10 of ImI, hence the ImI decreases 14,000-fold binding affinity to Ls-AChBP compared with Ac-AChBP [3,45]. Figure 5B shows that Leu-10 of TxIA (A10L) and Val-10 of BuIA [49] reside in positions similar to Leu-10 of PnIA (A10L D14K) and make contact with Arg-77, Val-106, Thr-108, Ser-112, and Met-114 of the Ac-AChBP complementary side, but Val-10 of BuIA forms fewer contacts due to the shorter length of its side-chain compared to Leu-10, resulting in a lower affinity to Ac-AChBP than PnIA (A10L D14K) [24].

α-Conotoxin Residues Selective for nAChRs
From the α-conotoxins complex with Ac-AChBP structure, and as shown in many experiments, it is assumed that α-conotoxin binding sites of nAChRs are similar to the binding sites of Ac-AChBP [1,36,49]. Sequence alignment of Ac-AChBP, Ls-AChBP, h(r)α3, h(r)α4, h(r)α6, h(r)α7, and h(r)α10 residues forming binding sites on the principal side are shown in Figure 6A. The residues forming binding sites on the complementary side are shown in Figure 6B. In Figure 6, it is shown that different nAChRs have different residues in the binding site, which could affect the conotoxin's affinity. For example, the Ile-57 (number 59 in that study [31]) of hα9 was replaced by Thr-57 in rα9 ( Figure 6B) result in α-conotoxin Vc1.1 increased affinity 10 folds for the rat receptor [31]. Modelling conotoxins binding to nAChRs could determine the key residues of α-conotoxins for a higher affinity to nAChRs [18,31,49]. Figure 6. Primary sequence alignment of Ac-AChBP, acetylcholine-binding protein from Lymnaea stagnalis (Ls-AChBP), h(r)α3, h(r)α4, h(r)α6, h(r)α7, h(r)α10, h(r) β2, h(r) β4, and h(r) α9 residues forming the binding sites to accommodate α-conotoxins. (A) The residues forming the binding sites on the principal side; (B) the residues forming the binding sites on the complementary side.

Residues of α-Conotoxin GIC Selective for hα3β2
As we have mentioned before, Gln-13 and His-5 of α-conotoxin GIC have high affinity for hα3β2. By docking α-conotoxin GIC to hα3β2, it was found that the GIC/hα3β2 complex was similar to the GIC/Ac-AChBP crystal structure. Mutation of His-5, Asn-11, and Asn-12 to Ala in α-conotoxin GIC ( Figure 7A) caused a loss of affinity for Ac-AChBP. The mutations also lost affinity for hα3β2. Only the Gln13-Ala substitution had the least influence on GIC activity in hα3β2 or Ac-AChBP. Primary sequence alignment of Ac-AChBP, acetylcholine-binding protein from Lymnaea stagnalis (Ls-AChBP), h(r)α3, h(r)α4, h(r)α6, h(r)α7, h(r)α10, h(r) β2, h(r) β4, and h(r) α9 residues forming the binding sites to accommodate α-conotoxins. (A) The residues forming the binding sites on the principal side; (B) the residues forming the binding sites on the complementary side.

Residues of α-Conotoxin GIC Selective for hα3β2
As we have mentioned before, Gln-13 and His-5 of α-conotoxin GIC have high affinity for hα3β2. By docking α-conotoxin GIC to hα3β2, it was found that the GIC/hα3β2 complex was similar to the GIC/Ac-AChBP crystal structure. Mutation of His-5, Asn-11, and Asn-12 to Ala in α-conotoxin GIC ( Figure 7A) caused a loss of affinity for Ac-AChBP. The mutations also lost affinity for hα3β2. Only the Gln13-Ala substitution had the least influence on GIC activity in hα3β2 or Ac-AChBP. From the GIC/hα3β2 complex model, it was found that Gln-13-Ala substitution had little effect, and it fit well with the receptor. Gln-13 or Ala-13 of GIC also fit well with the hα3β2 receptor, but Gln-13 of GIC caused a steric clash with Arg-108 in the hβ4 complementary binding side ( Figure 6B) and lost affinity to hα3β4 [35,47]. of GIC caused a steric clash with Arg-108 in the hβ4 complementary binding side ( Figure 6B) and lost affinity to hα3β4 [35,47].

Residues of α-Conotoxin PnIA Selective for hα7
As previously mentioned, A10L in PnIA (A10L) has higher affinity to Ac-AChBP and hα7 nAChR [36]. The docking study found that Leu-10 of PnIA (A10L) interacts with the hα7 nAChR hydrophobic pocket, which is composed of the residues Leu-106, Asn-108, His-112, and Gln-114 in the complementary side ( Figure 6B). Consequently, A10L mutation could extend the hydrophobic regions and increase the binding to the hydrophobic pocket of the receptor [3,36,37].
PnIA mutations ( Figure 7B) PnIA (L5R, A10L, D14R) display a 10-fold higher affinity for the hα7 nAChR compared to that of PnIA (A10L) [37]. The reason could be explained by Leu-5 of the PnIA being in the vicinity of hα7 residues , and Tyr-91. In the L5R mutation, the Arg-5 of the PnIA could form hydrogen bonds with those residues of hα7. Asp-14 of the PnIA is in the vicinity of hα7 residues His-114, Asn-110, and Lys75. The D14K mutation in PnIA (A10L) does not have much effect on its affinity to hα7 [36], but PnIA (L5R, A10L, D14R) shows high affinity to hα7; this may be due to the fact that the Arg-14 forms hydrogen bonds with Asn-110 residues of hα7 nAChRs [37].

Residues of α-Conotoxins SIA Selective for Torpedo nAChR
The PnIA (A10L D14K) mutation shows higher affinity for Ls-AChBP. Modeling studies found that the Lys-14 residue of PnIA (A10L D14K) forms a salt bridge with Glu-110 of the Ls-AChBP complementary site [36]. Kasheverov et al. found that adding positively charged residues (Lys or Arg) in the C-terminal region of α-conotoxins increased binding affinity to some nAChRs [15]. For example, the α-conotoxin SIA (D12K) mutation ( Figure 7C) showed higher affinity of Torpedo nAChRs. The reason could be similar to the D14K mutation in PnIA (A10L) which gained a higher affinity for Ls-AChBP [15]. Its Lys-12 in SIA (D12K) formed a salt bridge with Glu-57 of the Torpedo nAChRγ subunit and increased binding affinity to it [15].

Residues of α-Conotoxins MII Selective for rα6β2 nAChRs
Gotti et al. discovered that the α-conotoxin MII selectively blocked rα6β2 nAChRs [40], Gly-1 of MII was mutated to three amino acids RDP (arginine (R), aspartic acid (D), and proline (P)) ( Figure 7D), and the N-terminal amino acid tail motif RDP of MII increased affinity for the rα6β2 receptor by 13-fold. Modeling studies found that Arg-1 of the RDP motif makes hydrogen bonds with Asp-166 and Asp-167 of the β2 subunit [40].
Mutation of the conotoxin MII to MII (E11R) with a highly positively charged residue Arg-11 also favors local contact with the negatively charged area (Glu-19, Asp-148, Glu-151, Glu-191), and Arg-11 makes a hydrogen bond with Glu-151 ( Figure 8A). So the MII (E11R) mutation also has affinity to rα6β2. Glu-19 and Glu-151 in the α6 chain are replaced by Ala-19 and Lys-151 in the α3 receptor, thus making the α3 receptor binding site more positively charged and not favorable for Arg-11 of MII(E11R) to be bound by the receptor. Then, the MII (E11R) lost affinity for rα3β2 ( Figure 8B). So MII (E11R) could be selective for rα6β2 vs. rα3β2 [40].

Residues of α-Conotoxins Vc1.1 Selective for α9α10 nAChR
Vc1.1 is a conotoxin from Conus Victoriae ( Figure 7E). Mutations of Vc1.1 residues 5-7 and 11-15 to Ala, Lys, or Asp led to a significant decrease of functional activity at the α9α10 nAChR. The reason is that all these mutations can cause a disruption in the native structure that result in all the peptides unable to bind to the receptor, or alternatively affecting the specific amino acid interactions [19].
In contrast, substitutions at two other positions, Ser-4 and Asn-9, could improve the functional activity of Vc1.1 at the α9α10 nAChR. Replacement of Ser-4 by a positive residue (Arg or Lys), is more favorable for the potency at the rα9α10 and hα9rα10 nAChRs, but not for other receptors [19]. Vc1.1 mainly binds to α9α10 nAChR at the α10α9 pocket (α10 is principle site (+), and α9 is complementary site (−)). The receptor residues involved in interactions with Ser-4 of Vc1.1 are Asp-169, Asp-166, and Thr-32 of α9(−) [31]. Replacement of Ser-4 by Arg or Lys in Vc1.1, may favor the positive residue binding to those three residues, so increases potency at the α9α10 nAChR. 13-fold. Modeling studies found that Arg-1 of the RDP motif makes hydrogen bonds with Asp-166 and Asp-167 of the β2 subunit [40].
Mutation of the conotoxin MII to MII (E11R) with a highly positively charged residue Arg-11 also favors local contact with the negatively charged area (Glu-19, Asp-148, Glu-151, Glu-191), and Arg-11 makes a hydrogen bond with Glu-151 ( Figure 8A). So the MII (E11R) mutation also has affinity to rα6β2. Glu-19 and Glu-151 in the α6 chain are replaced by Ala-19 and Lys-151 in the α3 receptor, thus making the α3 receptor binding site more positively charged and not favorable for Arg-11 of MII(E11R) to be bound by the receptor. Then, the MII (E11R) lost affinity for rα3β2 ( Figure  8B). So MII (E11R) could be selective for rα6β2 vs. rα3β2 [40].

Residues of α-Conotoxins Vc1.1 Selective for α9α10 nAChR
Vc1.1 is a conotoxin from Conus Victoriae ( Figure 7E). Mutations of Vc1.1 residues 5-7 and 11-15 to Ala, Lys, or Asp led to a significant decrease of functional activity at the α9α10 nAChR. The reason is that all these mutations can cause a disruption in the native structure that result in all the peptides unable to bind to the receptor, or alternatively affecting the specific amino acid interactions [19].
In contrast, substitutions at two other positions, Ser-4 and Asn-9, could improve the functional activity of Vc1.1 at the α9α10 nAChR. Replacement of Ser-4 by a positive residue (Arg or Lys), is The receptor residues involved in interactions with Asn-9 of Vc1.1 are Gln-34, Arg-57, Thr-59, Thr-117 of rα9(−).The Thr-59 of rα9(−) is replaced by Ile-59 in hα9(−) (the number in that study). Asn-9, when replaced with a hydrophobic residue (Ala, Ile, Leu), will form more hydrophobic regions, so affinity to the receptors increased and resulted in more favorable binding potency at the rα9α10 and hα9rα10 nAChRs [19].
From those findings above, it could be suggested that the molecular structure combined with technology such as scanning mutagenesis, electrophysiology, etc. [18,19], not only helped us find the residues of α-conotoxins for various nAChR subtype selectivity, but also helped us design α-conotoxin analogs of higher potency and selectivity for drug application [41,[53][54][55].
For example, the α-conotoxin Vc1.1 potent variants such as N9A, N9I, N9L and S4R, which are highly selective for rα9α10, were first discovered by scanning mutagenesis by Craik et al. [19]. Using different computational strategies for modeling α-conotoxin Vc1.1 binding to nAChRs further determined key residues and binding sites that helped with the design of the new selective analog Vc1.1 (N9W) [31]. It also helps in the design of stable disulfide-deleted mutant of cyclic α-conotoxin Vc1.1 (C2H C8F) for oral drug application to treatment of neuropathic pain [56,57].

Conclusions
The diversity of α-conotoxins makes them not only excellent pharmacological probes for the dissection of structural features and functioning of nAChRs but also provides drug leads; some α-conotoxins are currently undergoing preclinical evaluation for the treatment of chronic and neuropathic pain. X-ray structures of α-conotoxins in complex with Ac-AChBP provide high-resolution information of these amino acid interactions and binding patterns. Based on co-crystal structures, computer models further revealed α-conotoxin residue selectivity for different nAChR subtypes and helped with the design of potent analogs. So, many potent α-conotoxins have been discovered by using the molecular structure of α-conotoxins selective for nAChRs, and here we only list some representatives. Currently, the optimization of drugs usually depends on animal models, but the drugs intended for human use could be misleading, as large activity differences can arise between binding site residues displaying some differences. A structural base provides insight to the important residues for binding to receptors and aids in the design of α-conotoxin analogs that are more potent for accurate treatment.