Spatial Structure and Activity of Synthetic Fragments of Lynx1 and of Nicotinic Receptor Loop C Models

Lynx1, membrane-bound protein co-localized with the nicotinic acetylcholine receptors (nAChRs) and regulates their function, is a three-finger protein (TFP) made of three β-structural loops, similarly to snake venom α-neurotoxin TFPs. Since the central loop II of α-neurotoxins is involved in binding to nAChRs, we have recently synthesized the fragments of Lynx1 central loop, including those with the disulfide between Cys residues introduced at N- and C-termini, some of them inhibiting muscle-type nAChR similarly to the whole-size water-soluble Lynx1 (ws-Lynx1). Literature shows that the main fragment interacting with TFPs is the C-loop of both nAChRs and acetylcholine binding proteins (AChBPs) while some ligand-binding capacity is preserved by analogs of this loop, for example, by high-affinity peptide HAP. Here we analyzed the structural organization of these peptide models of ligands and receptors and its role in binding. Thus, fragments of Lynx1 loop II, loop C from the Lymnaea stagnalis AChBP and HAP were synthesized in linear and Cys-cyclized forms and structurally (CD and NMR) and functionally (radioligand assay on Torpedo nAChR) characterized. Connecting the C- and N-termini by disulfide in the ws-Lynx1 fragment stabilized its conformation which became similar to the loop II within the 1H-NMR structure of ws-Lynx1, the activity being higher than for starting linear fragment but lower than for peptide with free cysteines. Introduced disulfides did not considerably change the structure of HAP and of loop C fragments, the former preserving high affinity for α-bungarotoxin, while, surprisingly, no binding was detected with loop C and its analogs.


Introduction
α-Neurotoxins from the snake venoms having a three-loop (or three-finger) spatial structure (three-finger proteins, TFPs) are well-recognized tools in the research on the nicotinic acetylcholine receptors (nAChRs) [1][2][3]. In humans and other organisms, there are also TFPs of the Ly6 family: these proteins are either water-soluble or attached by the glycosylphosphatidyinositol anchor to the membranes and some of them are targeting Thus, we were interested what effects might be introduced by a fixing disulfide on the structure and activity of the fragments both of the cholinergic ligand and of the ligandbinding site. It was found that the linear form of the synthetic fragments of TFPs of the Ly6 family is disordered, while the disulfide-cyclized form has a more stable conformation and is very similar to the respective portion of loop II within the ws-Lynx1 [26]. However, this property does not strongly affect the ability of both forms to bind to the nAChR from T. californica. The disulfide in loop C of AChBP had slightly increased the stability of the peptide but no ordered structure and no toxin binding was found. The additional disulfide did not strongly change a quite high secondary structure in HAP and did not affect its high affinity for α-Bgt.

Synthesis of Linear and Disulfide-Stabilized Fragments of Human Lynx1, of the Lymnaea Stagnalis AChBP Loop C and of the Combinatorial Peptide HAP
Linear peptides were prepared by solid-phase peptide synthesis using Fmoc/t-butyl strategy on tritylchloride-polystyrene resin (Intavis, Cologne, Germany) mostly as in [10]. Polypeptide chain assembly was performed on Syro I automatic peptide synthesizer (Mul-tiSynTech AG, Witten, Germany). Disulfides formation was carried out under conventional oxidation conditions: prolonged incubation on air in 50% aqueous acetonitrile at room temperature in the presence N-ethyldiisopropylamine (pH 8.0). In the case of formation of vicinal disulfide in peptide Ls202-214 C-C , the reaction proceeded extremely slowly with predominant formation of by-products, the best yields of target product being achieved by adding 20% dimethylsulfoxide to the peptide solution. All peptides were purified by highperformance liquid chromatography (HPLC) on a 250 × 30 mm C8 column (Dr. Maisch, Ammerbuch-Entringen, Germany) in linear gradient of acetonitrile from 10% to 40% and then lyophilized. The peptides purity was confirmed using analytical HPLC and the respective molecular masses were determined by MALDI-MS using Bruker Ultraflex I mass-spectrometer (Bruker Daltonik GmbH, Bremen, Germany) ( Table 1).

Activity of Synthetic Fragments of Lynx1 and of Nicotinic Receptor Loop C Models in Radioligand Assay
In these studies, we used radiolabelled α-Bgt ( 125 I-α-Bgt) with specific radioactivity of 500 Ci/mmol prepared as described in [27]. Competition experiments with Lynx1 fragments (compounds (1)-(3)) were performed mostly as described in [10]. Briefly, the membranes from T. californica ray electric organ (0.6 nM of α-bungarotoxin binding sites) were incubated in 50 µL of binding buffer (20 mM Tris-HCl buffer, pH 8.0, containing 1 mg/mL BSA) for 90 min with different concentrations of the Lynx1 peptides, followed by an additional 5-min incubation with 0.5 nM of 125 I-α-Bgt. The membranes were applied to glass GF/C filters (Whatman, Maidstone, UK) pretreated with 0.3% polyethylenimine. The samples were then washed (3 × 4 mL) with cold 20 mM Tris-HCl buffer, pH 8.0, containing 0.1 mg/mL BSA and bound radioactivity was measured with a Wallac 1470 Wizard Gamma Counter (PerkinElmer, Waltham, MA, USA). Nonspecific 125 I-α-Bgt binding was determined in the presence of 200-fold excess of α-cobratoxin. The results were analyzed using ORIGIN 7.5 (OriginLab, Northampton, MA, USA) fitting to one-site dose-response competition curves.
For the loop C fragments, HAP and its analogs (compounds (4)-(7)) we applied the same radioligand assay protocol, although in this case the observed decrease in binding of the 125 I-α-Bgt to the T. californica nAChR is caused not by direct competition of radioligand with the studied compounds (representing in this case the receptor model), but by their interaction with the 125 I-α-Bgt itself. In these experiments we incubated the Torpedo membranes (0.6 nM of α-bungarotoxin binding sites) for 180 min with 10 µM of the studied peptides, followed by an additional 5-min incubation with 0.5 nM of 125 I-α-Bgt and similar washing and counting.

Activity of Synthetic Fragments of Lynx1 and of Nicotinic Receptor Loop C Models in Radioligand Assay
In these studies, we used radiolabelled α-Bgt ( 125 I-α-Bgt) with specific radioactivity of 500 Ci/mmol prepared as described in [27]. Competition experiments with Lynx1 fragments (compounds (1)-(3)) were performed mostly as described in [10]. Briefly, the membranes from T. californica ray electric organ (0.6 nM of α-bungarotoxin binding sites) were incubated in 50 μL of binding buffer (20 mM Tris-HCl buffer, pH 8.0, containing 1 mg/mL BSA) for 90 min with different concentrations of the Lynx1 peptides, followed by an additional 5-min incubation with 0.5 nM of 125 I-α-Bgt. The membranes were applied to glass GF/C filters (Whatman, Maidstone, UK) pretreated with 0.3% polyethylenimine. The samples were then washed ( Combinatorial high-affinity peptide HAP with cysteines at N-and Ctermini, cyclized form. (Loop C), linear form.
Combinatorial high-affinity peptide HAP with cysteines at Nand C-termini, cyclized form.

Activity of Synthetic Fragments of Lynx1 and of Nicotinic Receptor Loop C Models in Radioligand Assay
In these studies, we used radiolabelled α-Bgt ( 125 I-α-Bgt) with specific radioactivity of 500 Ci/mmol prepared as described in [27]. Competition experiments with Lynx1 fragments (compounds (1)-(3)) were performed mostly as described in [10]. Briefly, the membranes from T. californica ray electric organ (0.6 nM of α-bungarotoxin binding sites) were incubated in 50 μL of binding buffer (20 mM Tris-HCl buffer, pH 8.0, containing 1 mg/mL BSA) for 90 min with different concentrations of the Lynx1 peptides, followed by an additional 5-min incubation with 0.5 nM of 125 I-α-Bgt. The membranes were applied to glass GF/C filters (Whatman, Maidstone, UK) pretreated with 0.3% polyethylenimine. The samples were then washed (3 × 4 mL) with cold 20 mM Tris-HCl buffer, pH 8.0, containing 0.1 mg/mL BSA and bound radioactivity was measured with a Wallac 1470 Wizard Gamma Counter (PerkinElmer, Waltham, MA, USA). Nonspecific 125 I-α-Bgt binding was determined in the presence of 200-fold excess of α-cobratoxin. The results were analyzed using 1892.8/1892. 9 HAP C-C

CD Analysis of the Synthetic Peptides
CD spectra were recorded using a JASCO J-810 spectropolarimeter (JASCO International Co., Tokyo, Japan) in the 190-250 nm spectral range in a 0.01 cm cuvette at 22 • C. Measurements were performed in the 50 mM sodium phosphate buffer (pH 7.5) at concentrations of 0.55, 0.44, 0.49, 0.68, 0.18, 0.73, 0.54, 0.59, 0.53 and 0.44 mM for compounds (1)-(10), respectively. Spectra were averaged over 4 consequent scans for each peptide. The secondary structure of peptides was calculated with the CONTIN/LL program (CDPro package) using the SMP56 protein reference set.
To assess the backbone dynamics of the peptides, the steady-state 1 H, 13 C nuclear Overhauser effect (NOE) and 13 C longitudinal relaxation rates (R1) were measured for all available Cα-H groups. Obtained values were converted to the order parameters of Cα-H vectors (S 2 ) and characteristic times of internal motions τ i using the formalism described in the works [32,33]. Overall rotational diffusion correlation times and hydrodynamic radii of the peptides were assessed using the Stokes-Einstein relationship from the coefficients of translational diffusion, D corrected for the peptide concentration [34]. The latter were measured by NMR using the PGSTE-watergate pulse sequence with simultaneous suppression of the convection effects and solvent signal [35].

Synthesis of Linear and Disulfide-Stabilized Fragments of Human Lynx1 Loop II, of the L. stagnalis AChBP Loop C and of the Combinatorial Peptide HAP
All linear peptides were synthesized using the standard Fmoc-strategy. To obtain the target peptides Ly29-43 C-C , Ls202-214 C-C , mLs202-214 C-C and HAP C-C , the disulfide bonds formation was performed by oxidation with air oxygen, giving a final yield of cyclized peptides at this stage from 30 to 40%, and did not cause difficulties with the exception of Ls202-214 C-C . In the case of Ls202-214 C-C product, even under the optimally selected conditions, the yield of the target peptide with vicinal disulfide did not exceed 3%. According to analytical HPLC data, the purity of the obtained target peptides was >95% and their masses closely corresponded to the calculated ones (Table 1). In order to prevent the formation of disulfides in peptides Ly29-43 C,C , mLs202-214 C,C and HAP C,C , they were stored always in the lyophilized form and dissolved immediately before the experiments.

Activity of Synthetic Fragments of Lynx1 and of Nicotinic Receptor Loop C Models in Radioligand Assay
The biological activity of Lynx1 fragments Ly29-43, Ly29-43 C,C and Ly29-43 C-C (peptides (1)-(3), respectively) was evaluated as their ability to compete with radioactive α-bungarotoxin ( 125 I-α-Bgt) for binding to the muscle-type nAChR from T. californica ray electric organ. All of them displaced the radioligand from the receptor in the range of µM concentrations ( Figure 1A). Herewith the linear peptide Ly29-43 and cyclized Ly29-43 C-C showed approximately equal inhibition efficiency (the respective IC 50 values were 7.9 ± 0.2 and 6.0 ± 0.2 µM). Interestingly, the linear form of Lynx1 fragment 29-43 with free cysteines at N-and C-termini was the most efficient with IC 50 = 1.1 ± 0.1 µM, with the activity considerably surpassing that of the ws-Lynx1.
The same competitive radioligand assay was used to evaluate the binding activity of L. stagnalis AchBP loop C and HAP analogs, although in this case, not the degree of competition between 125 I-α-Bgt and the tested peptides for binding to the target (T. californica nAChR) was evaluated, but the ability of these peptides to bind the radioligand itself, thus preventing its association with the receptor. This method allowed us to demonstrate the high efficiency of HAP analogs; at 10 µM they have completely bound 125 I-α-Bgt ( Figure 1B). The effectiveness of this binding did not depend on whether HAP was in a linear (peptides (8), (9)) or cyclized form (peptide (10)). Interestingly, the native L. stagnalis AChBP loop C (which served as the basis for the design of HAP) in both linear and cyclized forms (peptides (4)-(7)) did not reveal any ability to bind the radioligand at the same concentration ( Figure 1B). bungarotoxin ( 125 I-α-Bgt) for binding to the muscle-type nAChR from T. californica ra electric organ. All of them displaced the radioligand from the receptor in the range of μM concentrations ( Figure 1A). Herewith the linear peptide Ly29-43 and cyclized Ly29-43 C showed approximately equal inhibition efficiency (the respective IC50 values were 7.9 0.2 and 6.0 ± 0.2 μM). Interestingly, the linear form of Lynx1 fragment 29-43 with fre cysteines at N-and C-termini was the most efficient with IC50 = 1.1 ± 0.1 μM, with th activity considerably surpassing that of the ws-Lynx1.  (1)-(3), respectively. Each point is the mean ± SEM value of two measurements for each concentration in two independent experiments. The calculated IC50 values (mean ± SEM) for peptides 1-3 were 7.9 ± 0.2, 1.1 ± 0.1 and 6.0 ± 0.2 μM, respectively. For comparison, the inhibition curve for ws-Lynx1 (no symbols) on the same receptor under the same conditions (IC50 24 μM) is shown [26]. (B) Bar graph presentation of the decrease in specific binding of 125 I-α-Bgt to T. californica nAChR in the presence of L. stagnalis AChBP loop C and HAP analogs. Concentration of peptides (4)-(10) was 10 μM at which no "inhibition" for loop C fragments (102 ± 7 % of specific binding) and complete "inhibition" for HAP analogs (0.6 ± 0.1 % of specific binding) was revealed. The specific binding in the absence of peptides is accepted as 100%. Data are presented as mean ± SEM.
The same competitive radioligand assay was used to evaluate the binding activity o L. stagnalis AchBP loop C and HAP analogs, although in this case, not the degree of com petition between 125 I-α-Bgt and the tested peptides for binding to the target (T. californic nAChR) was evaluated, but the ability of these peptides to bind the radioligand itself, thu preventing its association with the receptor. This method allowed us to demonstrate th high efficiency of HAP analogs; at 10 μM they have completely bound 125 I-α-Bgt (Figur 1B). The effectiveness of this binding did not depend on whether HAP was in a linea (peptides (8), (9)) or cyclized form (peptide (10)). Interestingly, the native L. stagnali AChBP loop C (which served as the basis for the design of HAP) in both linear and cy clized forms (peptides (4)- (7)) did not reveal any ability to bind the radioligand at th same concentration ( Figure 1B).  (1)-(3), respectively. Each point is the mean ± SEM value of two measurements for each concentration in two independent experiments. The calculated IC 50 values (mean ± SEM) for peptides 1-3 were 7.9 ± 0.2, 1.1 ± 0.1 and 6.0 ± 0.2 µM, respectively. For comparison, the inhibition curve for ws-Lynx1 (no symbols) on the same receptor under the same conditions (IC 50 24 µM) is shown [26]. (B) Bar graph presentation of the decrease in specific binding of 125 I-α-Bgt to T. californica nAChR in the presence of L. stagnalis AChBP loop C and HAP analogs. Concentration of peptides (4)-(10) was 10 µM at which no "inhibition" for loop C fragments (102 ± 7 % of specific binding) and complete "inhibition" for HAP analogs (0.6 ± 0.1 % of specific binding) was revealed. The specific binding in the absence of peptides is accepted as 100%. Data are presented as mean ± SEM.

CD Analysis of the Synthetic Peptides
The obtained CD spectra and the content of secondary structure elements formally calculated from them allow us to make reliable conclusion about very low content of α-helices in the studied peptides in an aqueous solution (Table 2, Figure 2). The differences in content of other secondary structure elements, taking into account the relativity of these values for small peptides, are too small to draw certain conclusions.  [36], the analysis is considered to be reliable if NRMSD is < 0.1.

NMR Analysis of Synthetic Lynx1 Fragments
CD spectroscopy provides a quantitative view on the peptide secondary structure; however, the results lack the atomic details. Therefore, we applied the heteronuclear NMR spectroscopy to identify such atomic differences using the linear and cyclized forms of a Lynx1 loop II fragment: Ly29-43 C,C and Ly29--43 C-C which differed in their affinities towards T. californica nAChR ( Figure 1A). For both peptides we obtained the full 1 H, 13 C and 15 N chemical shift assignments, which were then analyzed using several approaches. As a first step, we characterized the overall compactness of the spatial structure by measuring the translational diffusion of the two peptides. The diffusion coefficient of Ly29-43 C-C was equal to (261 ± 2) 10 −12 m 2 /s, corresponding to the hydrodynamic radius of 1.07 nm and molecular weight of a globular protein of 2.1 kDa [37]. Thus, Ly29-43 C-C is compact and behaves as a globular protein, if this term is applicable for such a small peptide. In contrast, diffusion of Ly29-43 C,C is much slower, (215 ± 2)·10 −12 m 2 /s, which corresponds to 3.8 kDa. Therefore, Ly29-43 C,C does not form a compact spatial structure and behaves as a disordered protein.
Next, we analyzed the dynamic structure of two peptides using the concepts of random-coil index (RCI) [28] and secondary structure propensity (SSP) [29]. RCI converts the chemical shifts into the backbone order parameter RCI-S 2 . Regions with RCI-S 2 greater

NMR Analysis of Synthetic Lynx1 Fragments
CD spectroscopy provides a quantitative view on the peptide secondary structure; however, the results lack the atomic details. Therefore, we applied the heteronuclear NMR spectroscopy to identify such atomic differences using the linear and cyclized forms of a Lynx1 loop II fragment: Ly29-43 C,C and Ly29-43 C-C which differed in their affinities towards T. californica nAChR ( Figure 1A). For both peptides we obtained the full 1 H, 13 C and 15 N chemical shift assignments, which were then analyzed using several approaches. As a first step, we characterized the overall compactness of the spatial structure by measuring the translational diffusion of the two peptides. The diffusion coefficient of Ly29-43 C-C was equal to (261 ± 2) 10 −12 m 2 /s, corresponding to the hydrodynamic radius of 1.07 nm and molecular weight of a globular protein of 2.1 kDa [37]. Thus, Ly29-43 C-C is compact and behaves as a globular protein, if this term is applicable for such a small peptide. In contrast, diffusion of Ly29-43 C,C is much slower, (215 ± 2)·10 −12 m 2 /s, which corresponds to 3.8 kDa. Therefore, Ly29-43 C,C does not form a compact spatial structure and behaves as a disordered protein.
Next, we analyzed the dynamic structure of two peptides using the concepts of random-coil index (RCI) [28] and secondary structure propensity (SSP) [29]. RCI converts the chemical shifts into the backbone order parameter RCI-S 2 . Regions with RCI-S 2 greater than 0.7 might be considered as stable and structured, while lower magnitudes of the parameter correspond to the dynamic and partially disordered backbone. Ly29-43 C,C is characterized by low RCI-S 2 , 0.5 on average, reaching 0.6 close to the central Pro residue ( Figure 3A). In contrast, Ly29-43 C-C RCI-S 2 reveals a plateau of 0.7 on the region 4-12, indicating that the cyclic peptide is substantially more stable, however, is still partially disordered.
SSP is a well-established approach to describe the structure of intrinsically disordered proteins and peptides. SSP values greater than zero indicate the propensity of the helical conformation, while the negative SSPs reveal the possibility of β-sheet or extended (in contrast to disordered) conformation to be formed in the region. SSPs of Ly29-43 C-C residues reveal a relatively high propensity (up to 50%) of the extended conformation in the N-and C-terminal regions with the unstructured central part of the peptide ( Figure 3B). Comparison with the structure of the full size Lynx1 central loop (green bars in Figure 3B) reveals that the extended conformation occurs at the same sites, where the β-structure is observed in Lynx1. than 0.7 might be considered as stable and structured, while lower magnitudes of the parameter correspond to the dynamic and partially disordered backbone. Ly29-43 C,C is characterized by low RCI-S 2 , 0.5 on average, reaching 0.6 close to the central Pro residue (Figure 3A). In contrast, Ly29-43 C-C RCI-S 2 reveals a plateau of 0.7 on the region 4-12, indicating that the cyclic peptide is substantially more stable, however, is still partially disordered.  To further characterize the structure of Ly29-43 C-C , we applied the nuclear Overhauser effect NMR spectroscopy (NOESY). No characteristics for β-sheet backbone connectivities are observed in the NOESY NMR spectra, mainly due to the low chemical shift dispersion. However, two long-range side-chain contacts are seen that support the transient β-hairpin structure of Ly29-43 C-C (between the Y7 aromatic ring and V14 methyl groups and between the Y6 ring and K13 methylenes). All aforesaid implies that at least part of the time, Ly29-43 C-C resides in the β-hairpin conformation, close to structure of the homologous region of the parent Lynx1 protein. In contract, Ly29-43 C,C reveals the small 10-20% propensity for the extended conformation on the whole length of the peptide, which is far from the structure of this region within the full-length Lynx1.
Last, we employed the NMR relaxation parameters of α-carbons to assess the intramolecular mobility of the peptides' backbone. Two parameters, 1 H-13 C steady-state NOE (η) and 13 C longitudinal relaxation rate R1 are sufficient to measure the order parameters (S 2 ) and characteristic times, describing the internal motions [32]. NOEs of Ly29-43 C,C are substantially higher than those of Ly29-43 C-C , especially on the terminal regions of the peptide, implying the enhanced mobility ( Figure 3C). In turn, the detailed analysis of Ly29-43 C-C reveals that the peptide is relatively stable. We obtained an almost uniform S 2 distribution for all the residues (0.7-0.8) in this peptide (Figure 4).
is observed in Lynx1.
To further characterize the structure of Ly29-43 C-C , we applied the nuclear Overhauser effect NMR spectroscopy (NOESY). No characteristics for β-sheet backbone connectivities are observed in the NOESY NMR spectra, mainly due to the low chemical shift dispersion. However, two long-range side-chain contacts are seen that support the transient β-hairpin structure of Ly29-43 C-C (between the Y7 aromatic ring and V14 methyl groups and between the Y6 ring and K13 methylenes). All aforesaid implies that at least part of the time, Ly29-43 C-C resides in the β-hairpin conformation, close to structure of the homologous region of the parent Lynx1 protein. In contract, Ly29-43 C,C reveals the small 10-20% propensity for the extended conformation on the whole length of the peptide, which is far from the structure of this region within the full-length Lynx1.
Last, we employed the NMR relaxation parameters of α-carbons to assess the intramolecular mobility of the peptides' backbone. Two parameters, 1 H-13 C steady-state NOE (η) and 13 C longitudinal relaxation rate R1 are sufficient to measure the order parameters (S 2 ) and characteristic times, describing the internal motions [32]. NOEs of Ly29-43 C,C are substantially higher than those of Ly29-43 C-C , especially on the terminal regions of the peptide, implying the enhanced mobility ( Figure 3C). In turn, the detailed analysis of Ly29-43 C-C reveals that the peptide is relatively stable. We obtained an almost uniform S 2 distribution for all the residues (0.7-0.8) in this peptide (Figure 4).

NMR Analysis of Synthetic Analogs of the L. Stagnalis AChBP Loop C
Using a similar approach, we analyzed the properties of mutated cyclic and linear L. stagnalis AChBP loop C fragments-mLs202-214 C-C and mLs202-214 C,C . Similarly to the Lynx1 loop, the linear peptide appeared less compact: (282 ± 5)·10 −12 m 2 /s and 1.7 kDa vs. (256 ± 3)·10 −12 m 2 /s and 2.2 kDa. The linear peptide mLs202-214 C,C is also much more mobile than cyclic mLs202-214 C-C , according to the S 2 magnitudes ( Figure 5C). On the other hand, unlike Lynx1, the effect of cyclization on the structure of the loop C peptides appeared rather subtle. RCI-S 2 profiles of mLs202-214 C,C and mLs202-214 C-C are highly similar and hardly approach 0.5, suggesting the dynamic behavior of two peptides. Both mLs202-214 C,C and mLs202-214 C-C are mostly unstructured with SSPs in the range −0.2: +0.1 ( Figure 5A,B). The cyclic peptide reveals a higher but still subtle propensity for the

NMR Analysis of Synthetic Analogs of the L. stagnalis AChBP Loop C
Using a similar approach, we analyzed the properties of mutated cyclic and linear L. stagnalis AChBP loop C fragments-mLs202-214 C-C and mLs202-214 C,C . Similarly to the Lynx1 loop, the linear peptide appeared less compact: (282 ± 5)·10 −12 m 2 /s and 1.7 kDa vs. (256 ± 3)·10 −12 m 2 /s and 2.2 kDa. The linear peptide mLs202-214 C,C is also much more mobile than cyclic mLs202-214 C-C , according to the S 2 magnitudes ( Figure 5C). On the other hand, unlike Lynx1, the effect of cyclization on the structure of the loop C peptides appeared rather subtle. RCI-S 2 profiles of mLs202-214 C,C and mLs202-214 C-C are highly similar and hardly approach 0.5, suggesting the dynamic behavior of two peptides. Both mLs202-214 C,C and mLs202-214 C-C are mostly unstructured with SSPs in the range −0.2: +0.1 ( Figure 5A,B). The cyclic peptide reveals a higher but still subtle propensity for the extended conformation. No long-range NOESY cross-peaks were found in NMR spectra, in agreement with all data above. Thus, none of the peptides reveal the secondary structure of a parent protein ( Figure 5B) found in the respective X-ray structures of the parent AChBP. extended conformation. No long-range NOESY cross-peaks were found in NMR spectra, in agreement with all data above. Thus, none of the peptides reveal the secondary structure of a parent protein ( Figure 5B) found in the respective X-ray structures of the parent AChBP.

Discussion
The present communication is devoted to the analysis of the TFPs interactions with nAChRs, but covers it at a "lower level", namely by studying the interactions between the TFP fragments and the whole-size nAChRs and, on the other hand, the associations between the nAChR binding site synthetic fragments or their models with the wholesize TFPs. To date, numerous biochemical studies have demonstrated that the main "beneficiaries" of high affinity of the most well-known snake α-neurotoxin TFPs towards distinct nAChR subtypes are the toxin central loop II and the loop C of the receptor αsubunit, respectively. Direct confirmation of this was obtained recently by a cryo-electron microscopy high resolution structure of the α-Bgt complex with the Torpedo nAChR [18] ( Figure 6A), while earlier the relevant information was based on the X-ray structures of α-neurotoxins in complexes with AChBPs [20,23] (Figure 6B) and with the monomeric ligand-binding domains of the α1 and α9 subunits [19,24].

Discussion
The present communication is devoted to the analysis of the TFPs interactions with nAChRs, but covers it at a "lower level", namely by studying the interactions between the TFP fragments and the whole-size nAChRs and, on the other hand, the associations between the nAChR binding site synthetic fragments or their models with the whole-size TFPs. To date, numerous biochemical studies have demonstrated that the main "beneficiaries" of high affinity of the most well-known snake α-neurotoxin TFPs towards distinct nAChR subtypes are the toxin central loop II and the loop C of the receptor α-subunit, respectively. Direct confirmation of this was obtained recently by a cryo-electron microscopy high resolution structure of the α-Bgt complex with the Torpedo nAChR [18] ( Figure  6A), while earlier the relevant information was based on the X-ray structures of α-neurotoxins in complexes with AChBPs [20,23] (Figure 6B) and with the monomeric ligandbinding domains of the α1 and α9 subunits [19,24]. These X-ray and cryo-electron microscopy data are extremely important both for fundamental research and for practical purposes. In particular, our recent paper [10] was initiated by the well-known facts that synthetic fragments of the central loop II of some snake venom TFPs possess a partial activity of the starting neurotoxins against nAChRs, especially when their structure is stabilized by the two disulfide-joined Cys residues grafted at the N-and C-termini (conventionally such peptides can be named as "cyclized") (see, for example review [7]). The novelty of that communication was the synthesis and analysis of binding of such fragments of those TFPs of the Ly6 family which are endogenous regulators of the nAChRs in human and other organisms. Interestingly, against the These X-ray and cryo-electron microscopy data are extremely important both for fundamental research and for practical purposes. In particular, our recent paper [10] was initiated by the well-known facts that synthetic fragments of the central loop II of some snake venom TFPs possess a partial activity of the starting neurotoxins against nAChRs, especially when their structure is stabilized by the two disulfide-joined Cys residues grafted at the N-and C-termini (conventionally such peptides can be named as "cyclized") (see, for example review [7]). The novelty of that communication was the synthesis and analysis of binding of such fragments of those TFPs of the Ly6 family which are endogenous regulators of the nAChRs in human and other organisms. Interestingly, against the T. californica nAChR, the fragment of the water-soluble Lynx1 revealed the activity quite close to that of the starting protein. Another our paper [38] was devoted to a short combinatorial peptide HAP which is homologous to the nAChR loop C and binds α-Bgt with a surprisingly high affinity (2-4 nM), which is virtually the same as for α-Bgt binding to the whole-size nAChRs [15]. Kudryavtsev et al. [38] confirmed this result with a considerably extended arsenal of methods for the analysis of HAP binding. After these publications we had several questions: what determines the activity of the synthetic fragments of Lynx1 and of loop C models, what is the real structure of these peptides, how close they are to the conformation of their parent ligands and receptors and what is the role of disulfides?
To answer these questions, we synthesized an additional set of fragments. Keeping in mind that the disulfide introduction at the peptide termini of the TFP loop II in several cases increased the affinity for the nAChRs, we decided to apply such an approach to the synthetic receptor models (HAP and fragment of loop C from the L. stagnalis AChBP). That is why we synthesized the analogs of fragments of the Lynx1 central loop II, the loop C from L. stagnalis AChBP and the high-affinity peptide HAP, which spatial structures were stabilized by a disulfide bond formed between the additional cysteine residues introduced at the N-and C-termini of these fragments: Ly29-43 C-C , mLs202-214 C-C and HAP C-C , respectively. The control samples in the biological tests were their linear (non-oxidized) forms-Ly29-43 C,C , mLs202-214 C,C and HAP C,C , as well as two original fragments without flanking cysteines-Ly29-43 and HAP. In the case of the loop C fragment, its mutant form was also used with the replacement of vicinal cysteines with serines to avoid the formation of mis-folded Cys-isomers. In addition, it was previously shown that the presence of vicinal disulfide in the corresponding fragments of the nAChR α1-subunit does not strongly affect the binding of snake venom neurotoxins [39,40]. The natural loop C with vicinal cysteines closed in disulfide (Ls202-214 C-C ) or in linear form (Ls202-214 C,C ) were characterized for their activity as well. The chemical structures of all synthesized peptides were confirmed by mass spectrometry (Table 1) and their biological activities were evaluated by competitive radioligand assay on T. californica nAChR.
The biological activity of fragments of several other Ly6 proteins, namely, their ability to compete with radioactive α-Bgt for binding to different nAChRs and such their models as AChBPs and ligand-binding domains of individual subunits, as well as the ability of these fragments to inhibit currents in certain nAChR subtypes, was described by us in detail earlier [10]. The present study confirmed the micromolar affinity of fragments of the central loop II of the Lynx1 protein towards the T. californica nAChR (IC 50 values were in the range 1.1-7.9 µM), showing that the inhibitory activity was higher than that of the full-size protein for the same receptor or AChBPs from L. stagnalis and Aplysia californica [25] and does not strongly depend on whether the fragment is linear or cyclized ( Figure 1A).
These AChBPs, which are structural homologues of ligand-binding domains of nAChRs and often reproduce the nAChR pharmacological profiles, are still actively used in the X-ray crystallography in complexes with various cholinergic ligands. After these structural studies, it became possible to clarify the mechanism of nAChRs functioning and the detailed spatial structure of their orthosteric binding sites, although information about the key amino acid residues of the receptor binding site already existed. In particular, the important role of loop C of the α-subunits of nAChRs was known, which was used for the combinatorial design of a short 13-member peptide that mimics this fragment of the T. californica receptor. This peptide was called a high-affinity peptide (HAP) because it showed a nanomolar affinity for α-Bgt, a blocker of distinct nAChR subtypes [15], and it mimicked the structure of the receptor loop C in complex with this toxin [16]. Interestingly, the HAP sequence combines the amino acid residues of the α-subunit of T. californica nAChR and L. stagnalis AChBP.
In the present study, for the first time, we intended to evaluate the influence of spatial structure on the activity of the L. stagnalis AChBP loop C fragment in linear and cyclized forms and to compare it with those for HAP and its analogs. The introduction of additional flanking cysteines into the HAP sequence and their cyclization did not affect the high ability of this peptide to bind 125 I-α-Bgt, thus preventing association of this radioligand with T. californica nAChR ( Figure 1B). To our surprise, neither linear nor cyclized forms of loop C where two vicinal Cys were substituted for Ser (as in HAP) could bind the radioligand. This result cannot be explained by the mutation introduced into the loop C structure, since a similar lack of ability to bind 125 I-α-Bgt was detected for the fragment with vicinal disulfide ( Figure 1B).
This difference in biological activity between the native loop C and its combinatorial variant-HAP-may be due to the difference in their spatial structure. Therefore, we estimated the content of secondary structure elements by circular dichroism for all synthesized peptides. The application of CD to small peptide molecules to characterize their secondary structure may raise questions, but there is a sufficient number of examples in the literature when a disruption, for example, of the disulfide bond in a small peptide leads to a noticeable change in the CD spectrum [41][42][43]. We expected that the introduction of two Cys at the ends of the peptide termini with the subsequent closure of the disulfide could affect the content of secondary structure elements and possibly lead to stabilization of the spatial structure. However, in our case the CD measurements did not reveal substantial difference between the linear and cyclized forms ( Figure 2, Table 2).
In order to obtain more precise information about the spatial structure and stability of peptides under investigation, we applied the solution NMR approaches to four objects: Ly29-43 C,C , Ly29-43 C-C , mLs202-214 C,C and mLs202-214 C-C . All the peptides under investigation were partially disordered, thus we had to utilize the secondary structure propensity techniques and NMR relaxation analysis to describe the structural ensembles present in solution. Among all the peptides, the ws-Lynx1 fragment fixed by a disulfide Ly29-43 C-C was the most stable, with SSP up to 50% in several regions, which allowed us to apply the conventional NMR approaches to solve the structure of the peptide in solution (Figure 7). Although being in general more mobile than the respective loop in the ws-Lynx1, the Ly29-43 C-C has a very similar conformation to the respective part of ws-Lynx1. forms and to compare it with those for HAP and its analogs. The introduction of add tional flanking cysteines into the HAP sequence and their cyclization did not affect th high ability of this peptide to bind 125 I-α-Bgt, thus preventing association of this radiolig and with T. californica nAChR ( Figure 1B). To our surprise, neither linear nor cyclize forms of loop C where two vicinal Cys were substituted for Ser (as in HAP) could bin the radioligand. This result cannot be explained by the mutation introduced into the loo C structure, since a similar lack of ability to bind 125 I-α-Bgt was detected for the fragmen with vicinal disulfide ( Figure 1B). This difference in biological activity between the native loop C and its combinatoria variant-HAP-may be due to the difference in their spatial structure. Therefore, we es timated the content of secondary structure elements by circular dichroism for all synthe sized peptides. The application of CD to small peptide molecules to characterize their sec ondary structure may raise questions, but there is a sufficient number of examples in th literature when a disruption, for example, of the disulfide bond in a small peptide lead to a noticeable change in the CD spectrum [41][42][43]. We expected that the introduction o two Cys at the ends of the peptide termini with the subsequent closure of the disulfid could affect the content of secondary structure elements and possibly lead to stabilizatio of the spatial structure. However, in our case the CD measurements did not reveal sub stantial difference between the linear and cyclized forms ( Figure 2, Table 2).
In order to obtain more precise information about the spatial structure and stabilit of peptides under investigation, we applied the solution NMR approaches to four object Ly29-43 C,C , Ly29-43 C-C , mLs202-214 C,C and mLs202-214 C-C . All the peptides under invest gation were partially disordered, thus we had to utilize the secondary structure propen sity techniques and NMR relaxation analysis to describe the structural ensembles presen in solution. Among all the peptides, the ws-Lynx1 fragment fixed by a disulfide Ly29-43 C was the most stable, with SSP up to 50% in several regions, which allowed us to appl the conventional NMR approaches to solve the structure of the peptide in solution (Figur 7). Although being in general more mobile than the respective loop in the ws-Lynx1, th Ly29-43 C-C has a very similar conformation to the respective part of ws-Lynx1.  On the contrary, the linear peptide Ly29-43 C,C was highly flexible and disordered. However, it revealed even higher nAChR-binding activity as compared to the cyclized variant ( Figure 1A). This is an unexpected result, implying the conformational selection mechanism of nAChR binding. In addition, it is impossible to ignore the presence in the linear form of two active sulfhydryl groups, which can form additional hydrogen bonds in the receptor binding site, or even participate in disulfide exchange with vicinal cysteines of the C-loop of the receptor α-subunit thereby increasing the affinity.
A markedly different result was obtained for the synthetic L. stagnalis AChBP loop C fragments in linear or cyclized forms (mLs202-214 C,C , mLs202-214 C-C ). According to 1 H-NMR analysis (see Figure 5) none of these peptides had a fixed three-dimensional structure. We had already mentioned that the X-ray structures of complexes of various agonists and antagonists with the AChBPs, where the loop C of the latter was always playing a major role and was used to model the interactions with the whole-size nAChRs. What concerns the interactions with α-neurotoxins, it was earlier demonstrated for some fragments covering the loop C of the α1 nAChR subunit and containing vicinal cysteines that toxin binding was not extremely sensitive to the state (free or oxidized) of these cysteines [39,40]. In the present communication the AChBP loop C fragments were synthesized and tested in their interaction with α-Bgt for the first time and we still cannot explain why they were inactive and why the stabilization of the loop C spatial structure by disulfide closure did not give any promising results.
Summarizing, we have shown the effect of disulfide on the structure and stability of peptide fragments in one case (for loop II of Lynx1), sometimes (for loop C of AChBP and for HAP) it is absent, while the structuring of the peptides in solution does not necessarily correlate with their activity: the most active was the linear peptide Ly29-43 C,C , that contrasts with the least active but the most structured ligand-the full-size ws-Lynx1. Anyway, because for various Ly6 proteins such as ws-Lynx1 or SLURP-1 are carried out versatile tests against cancer and neurodegenerative diseases [44,45] we believe that a set of obtained and characterized synthetic fragments can also be useful.

Conclusions
The main result is that a set of the synthetic fragments of the loop II of Ly6 TFP was prepared, including those where the additional Cys residues were introduced at the Nand C-termini and left free or oxidized to the disulfide. The inhibitory activity of these peptides was tested against Torpedo nAChR and some linear forms or those "cyclized" by a disulfide were found to be more active than the ws-Lynx1. By 1 H-NMR we found that the disulfide-fixed Lynx1 fragment has a three-dimensional structure very similar to the loop II within the whole-size ws-Lynx1. We also tried to apply such "cyclization" approach to the models of the nAChR binding sites, namely to AChBP loop C and HAP. The latter retained its binding activity in linear and cyclized forms, in contrast with the loop C which did not reveal for both forms either the binding capacities or an ordered secondary structure.