Snake Toxins Labeled by Green Fluorescent Protein or Its Synthetic Chromophore are New Probes for Nicotinic acetylcholine Receptors

Fluorescence can be exploited to monitor intermolecular interactions in real time and at a resolution up to a single molecule. It is a method of choice to study ligand-receptor interactions. However, at least one of the interacting molecules should possess good fluorescence characteristics, which can be achieved by the introduction of a fluorescent label. Gene constructs with green fluorescent protein (GFP) are widely used to follow the expression of the respective fusion proteins and monitor their function. Recently, a small synthetic analogue of GFP chromophore (p-HOBDI-BF2) was successfully used for tagging DNA molecules, so we decided to test its applicability as a potential fluorescent label for proteins and peptides. This was done on α-cobratoxin (α-CbTx), a three-finger protein used as a molecular marker of muscle-type, neuronal α7 and α9/α10 nicotinic acetylcholine receptors (nAChRs), as well as on azemiopsin, a linear peptide neurotoxin selectively inhibiting muscle-type nAChRs. An activated N-hydroxysuccinimide ester of p-HOBDI-BF2 was prepared and utilized for toxin labeling. For comparison we used a recombinant α-CbTx fused with a full-length GFP prepared by expression of a chimeric gene. The structure of modified toxins was confirmed by mass spectrometry and their activity was characterized by competition with iodinated α-bungarotoxin in radioligand assay with respective receptor preparations, as well as by thermophoresis. With the tested protein and peptide neurotoxins, introduction of the synthetic GFP chromophore induced considerably lower decrease in their affinity for the receptors as compared with full-length GFP attachment. The obtained fluorescent derivatives were used for nAChR visualization in tissue slices and cell cultures.

The history of TFTs began with the discovery of α-bungarotoxin (α-BgTx) in the venom of the many-banded krait Bungarus multicinctus about 60 years ago (Chang and Lee, 1963). Dozens of TFTs, such as α-cobratoxin (α-CbTx) from the monocled cobra Naja kaouthia (Karlsson et al., 1972), neurotoxin NTII from the Caspian cobra Naja oxiana, and LsIII from the black-banded sea krait Laticauda semifasciata, were further purified and characterized (Utkin, 2019;Tsetlin et al., 2021). Those toxins served as high-affinity and selective molecular probes per se, and their labeled derivatives are traditionally utilized as robust molecular reporter tools in channel research (Anderson and Cohen, 1974;Stiles, 1993;Kuzmenkov and Vassilevski, 2018;Kudryavtsev et al., 2020). Pioneer works, where tritiated and iodinated α-BgTx was used, introduced the methodology of nAChR purification, identification of binding sites, and determination of the receptors distribution in the brain (O'Brien et al., 1972;Silver and Billiar, 1976;Lukasiewicz and Bennett, 1978). TFTs with isotopic moieties are still widely used in radioligand binding assays for the detection of novel compounds and establishing their kinetic properties (Martin et al., 2007;Anderson, 2008). An alternative approach to develop toxin-based molecular markers is the utilization of fluorescent dyes. Fluorescent derivatives of α-BgTx were used for visualizing the distribution of nAChR in vertebrate skeletal muscle fibers, expression profiling and description of the clusterization process of diffusely distributed receptors (Ravdin and Axelrod, 1977;Stya and Axelrod, 1983;Anderson, 2008;Shelukhina et al., 2009). Simultaneously, fluorescently-modified NTII and α-CbTx were also successfully produced and applied for binding studies and mapping of nAChR (Tsetlin et al., 1979(Tsetlin et al., , 1982Kang and Maelicke, 1980).
Here, we extend the techniques of toxin labeling by the design and production of novel fluorescent probes for nAChR imaging. For these purposes we obtained three derivatives of well-studied toxins. The first marker is the recombinant protein eGFP-α-CbTx, which is a fusion of α-CbTx and the enhanced green fluorescent protein (eGFP). α-CbTx binds with high affinity and practically irreversibly to nAChRs; therefore, even after some decrease in affinity as a result of modification this toxin retains the capacity to bind to its receptor. Earlier, it was shown that attachment of a fluorescein label at Lys23 in the central loop of α-CbTx resulted in a fluorescently labeled derivative retaining nanomolar affinity to Torpedo nAChR (Johnson and Taylor, 1982). For quite a long time we have been using α-CbTx and its derivatives for the study of different receptors (Utkin et al., 1998;Kudryavtsev et al., 2015) and gained extensive experience in work with just this toxin. Moreover, we possess sufficient quantities of native α-CbTx for comparison. All of this contributed to our decision to choose α-CbTx for preparation of chimera with eGFP. We have already successfully utilized conjugation with eGFP to produce and characterize selective probes to voltage-gated potassium channels (Kuzmenkov et al., 2016a). Such method of labeling is aimed at obtaining monotagged derivatives and claims to dramatically decrease the costs of the final products (Kuzmenkov and Vassilevski, 2018). The second fluorescent probe is presented by the same neurotoxin covalently linked to eGFP chromophore-mimicking moiety (p-HOBDI-BF 2 ) (Baranov et al., 2012). This tag was previously applied only for labeling DNA molecules (Stakheev et al., 2018) and used in the preparation of an inhibitor for irreversible interactions with the active-site cysteine of human cathepsins (Frizler et al., 2013). Moreover, to design specific muscle-type nAChR reporter we utilized p-HOBDI-BF 2 moiety for the preparation of a mono-labeled fluorescent derivative of azemiopsin (Aze), a linear neurotoxic peptide earlier isolated from the Azemiops feae viper venom (Utkin et al., 2012).
Chymotrypsin was from Sigma and Lys-C protease was from BDH Chemicals (Sweden). Recombinant human enteropeptidase light chain was a gift from Dr. Marine Gasparian (Gasparian et al., 2003). Acetonitrile was from PanReac-AppliChem (ITW Reagents, Spain). Commercially available reagents were used for activated p-HOBDI-BF 2 synthesis without additional purification. All other reagents were of analytical or higher grade.
To construct an expression vector bearing the eGFP-α-CbTx gene, we used pET-28a (Novagen). The gene of α-CbTx including a stop codon was synthesized by Evrogen (Russia).

Construction of a Vector Harboring the eGFP-α-CbTx Gene
At the first step the pET-28a-eGFP vector was obtained by cloning the eGFP gene alone, which was amplified with eGFP-f and eGFPr primers from the plasmid pUC-eGFP, as we described previously (Kuzmenkov et al., 2016a) (Table 1). NdeI and BamHI sites were used for cloning. On the second step, a sequence encoding a short linker was synthesized from primers (Link-f and Link-r) and added at BamHI and EcoRI restriction sites. Finally, the gene of α-CbTx with a stop codon was amplified with CbTx-f and CbTx-r oligonucleotides and then subcloned in pET-28a-eGFP using EcoRI and SalI restriction sites. Additionally, α-CbTx gene was produced with modified human enteropeptidase site (DDDDR) at the N-terminus by changing CbTx-f to CbTx-f-EK primer. All oligonucleotides from which the full genes encoding eGFP-α-CbTx and eGFP-EK-α-CbTx were generated by PCR amplification are listed in Table 1. Correct cloning of the genes was confirmed by sequencing.

Production and Purification of eGFP-α-CbTx and eGFP-EK-α-CbTx
Escherichia coli BL21 (DE3) cells transformed with pET-28a-eGFP-α-CbTx and pET-28a-eGFP-EK-α-CbTx were cultured at 37°C in LB medium in the presence of 50 ng/ml kanamycin to the mid-log phase. Expression of the fluorescent chimera gene was induced by 0.5 mM IPTG and the culture was further incubated at 25°C for 20 h. Cells were harvested by centrifugation, disrupted by sonication, and the chimeric proteins were purified from the soluble fraction by affinity chromatography on a TALON Superflow resin (Clontech) following the manufacturer's protocol. Further purification was performed by size-exclusion chromatography on a TSK 2000SW column (7.5 × 600 mm, 12.5 nm pore size, 10 μm particle size; Toyo Soda Manufacturing) in phosphate-buffered saline (PBS), pH 7.4, at a flow rate of 0.5 ml/min. Expression and purification of the target proteins was monitored by SDS-PAGE. Concentration of the final eGFP-α-CbTx and eGFP-EKα-CbTx preparations was measured by absorption spectroscopy using ε (489 nm) 55,000 M −1 cm −1 .
eGFP-EK-α-CbTx Digestion by Human Enteropeptidase Light Chain eGFP-EK-α-CbTx chimera produced in bacteria was subjected to treatment by enteropeptidase. The recombinant protein was dissolved in 50 mM Tris-HCl (pH 8.0) to a concentration of 1 mg/ml. Protein cleavage with human enteropeptidase light chain (1 U of enzyme per 1 mg of substrate) was performed overnight (16 h) at 37°C. The mixture was separated by reversedphase HPLC on a Jupiter C5 column (4.6 × 250 mm; Phenomenex) in a linear gradient of acetonitrile concentration (0-60% in 60 min) in the presence of 0.1% trifluoroacetic acid, at a flow rate of 1 ml/min. Purified recombinant α-CbTx was assessed by MALDI mass-spectrometry.

Mass Spectrometry
Ultraflex TOF-TOF (Bruker Daltonik, Germany) spectrometer was used for MALDI mass spectrometry as described previously (Kuzmenkov et al., 2016b). 2,5-Dihydroxybenzoic acid (Sigma-Aldrich) was used as a matrix. Measurements were performed in the linear mode. Mass spectra were analyzed with the Data Analysis 4.3 and Data Analysis Viewer 4.3 software (Bruker).
High-resolution mass spectra of activated p-HOBDI-BF 2 were recorded on a TripleTOF 5600+ System (SCIEX) using electrospray ionization (ESI). The measurements were done in a positive ion mode (interface capillary voltage of 5500 V); mass range from m/z 50 to m/z 3,000; external or internal calibration was done with the ESI Tuning Mix (Agilent). A syringe injection was used for solutions in acetonitrile, methanol, or water (flow rate of 20 μl/min). Nitrogen was applied as a dry gas; interface temperature was set at 180°C.
ESI mass spectra of Aze fragments were recorded on an LTQ Orbitrap XL instrument (Thermo Fisher Scientific) equipped with HESI-II ion source. The acquisition was performed using full-scan FTMS mode at 30K resolution in a positive ions detection mode within the 700-2000 mass range.

Localization of Label with Chromato-Mass Spectrometry Analysis
Purified HOBDI-BF 2 -Aze conjugate was dissolved in 50 mM ammonium bicarbonate solution (pH 9.0) to a final concentration of 1 mg/ml, then subjected to digestion by chymotrypsin or Lys-C protease (1 : 25, enzyme: substrate, w/w) at 37°C for 24 h, followed by formic acid quenching (1% by vol). The digested samples were analyzed by LC-MS on a YMC-Triart C18 column (2.1 × 150 mm; YMC) in a linear gradient of acetonitrile (5-55%) at flow rate of 0.3 ml/min. The modified peptide fragments were detected by specific dye absorption at 415 nm.

Competitive Radioligand Assay
In competition experiments with 125 I-α-BgTx (500 Ci/mmol), all tested compounds (α-CbTx, HOBDI-BF 2 -α-CbTx, eGFP-α-CbTx, eGFP, Aze, and HOBDI-BF 2 -Aze derivatives) at different concentrations were pre-incubated for 2.5 h at room temperature with the great pond snail Lymnaea stagnalis acetylcholine binding protein (AChBP) (at the final concentration of 2.5 nM), or with GH 4 C 1 cells (final concentration of 0.4 nM of toxin-binding sites), or Torpedo californica electric organ membranes (final concentration of 0.5 nM of toxin-binding sites) in 50 μL of binding buffer (20 mM Tris-HCl, 1 mg/ml bovine serum albumin (BSA), pH 7.5). After that 125 I-α-BgTx was added to AChBP, GH 4 C 1 cells or membranes to a final concentration of 0.5 nM, and the mixtures were additionally incubated for 5 min. Binding was stopped by rapid filtration on double DE-81 filters (Whatman) pre-soaked in the binding buffer (for AChBP) or on GF/C filters (Whatman) pre-soaked in 0.25% polyethylenimine (for GH 4 C 1 cells or membranes), unbound radioactivity being removed from the filters by washout (3 × 3 ml) with the binding buffer. Nonspecific binding was determined in all cases using 2.5 h preincubation with 10 μM α-CbTx. The binding results were analyzed using OriginPro 2015 program (OriginLab Corporation) fitting to a one-site dose-response curve with the following equation: where n H is the slope factor (Hill coefficient). Data are presented as means ± standard errors of means (S.E.M.).

Thermophoresis
For the microscale thermophoresis (MST) experiments, both the fluorescent ligands (HOBDI-BF 2 -α-CbTx or eGFP-α-CbTx) and AChBP were diluted in PBS, pH 7.2, containing 0.05% Tween. Then, glass capillaries (NanoTemper Technologies, Germany) were filled with different concentrations of AChBP and with the same concentration of the fluorescent ligand. The MST experiments were performed with Monolith NT.115 (NanoTemper Technologies, Germany) according to the manufacturer's recommendations. Before each experiment, a pretest was carried out to check the fluorophore's stability and exclude the sorption of the fluorescent molecule on the capillary or aggregation during thermophoresis. Notably, no aggregation or adsorption was observed for either HOBDI-BF 2 -α-CbTx or eGFP-α-CbTx. The MST experiment was carried out at 40% MST power using a green light emitting diode for excitation. MST data were analyzed and binding parameters were calculated using MO. Affinity Analysis 2 Software (NanoTemper Technologies, Germany). Data points were fitted to a K d model equation according to built-in software recommendations.

Fluorescent Labeling of nAChRs in Cell Cultures and Tissue Slices
One day before the transfection, Neuro-2a cells were applied to coverslips and placed inside 35-mm cell culture dishes with 2 ml of Eagle's medium. Transfection medium contained cDNA of human α7 nAChR (3 mg/ml) and the Lipofectamine 2000 transfection reagent (Invitrogen, Thermo Fisher Scientific). The cDNA-containing solution was replaced by cell culture medium 6-8 h after the transfection. Cryostat 10 μm-thick sections of rat tongue were fixed with isopropanol for 10 min at 4°C, rinsed with PBS and distilled water, air dried for 1 h (all further procedures were carried out at room temperature) and incubated for 1 h with PBS, pH 7.4, containing 10 mg/ml BSA and 5 ml/L Tween 20 to block unspecific binding of toxins. The sections were pre-incubated for 1-2 h in buffer A (1 mg/ml BSA and 150 mM NaCl in PBS). Controls were run simultaneously by adding 300-fold excess of unlabeled α-CbTx (unless specified otherwise). Then, AF555-α-BgTx was added to the slides to reach 50 nM final concentration and they were incubated for 1-15 h. Sections were subsequently washed with PBS, fixed with 40 mg/ml paraformaldehyde for 10 min, rinsed with PBS again and coverslipped in carbonatebuffered glycerol at pH 8.6. The slides were analyzed using CellA Imaging Software (Olympus Soft Imaging Solutions, Germany) coupled to epifluorescent microscope with cooled CCD CAM-XM10 (Olympus, Japan).

Design and Production of Chimeric eGFP-α-CbTx
We designed an expression cassette encoding a three-finger toxin fused with a fluorescent protein via a flexible linker ( Figure 1A). The well-studied α-CbTx (UniProt ID: P01391) from the venom of N. kaouthia (Karlsson et al., 1972) was selected because this protein is highly active towards several targets including muscle-type, neuronal α7 and α9/α10 nAChRs (Chandna et al., 2019), as well as AChBP from L. stagnalis (Hansen et al., 2002); α-CbTx binds to these targets with sub-nanomolar affinities. Additionally, it was shown that α-CbTx is able to inhibit ionotropic γ-aminobutyric acid receptors (GABA A ) at sub-micromolar concentrations (Kudryavtsev et al., 2015). As a fluorescent module, we selected one of the best-studied fluorescent proteins eGFP (FPbase ID: R9NL8) (Cormack et al., 1996) which is characterized by high brightness and weak dimerization properties.
The eGFP-α-CbTx chimera was produced recombinantly, isolated from E. coli lysate by affinity chromatography, and purified by size-exclusion chromatography ( Figure 1B). Fluorescence excitation and emission spectra of this hybrid were found identical to those of native eGFP protein (excitation and emission maxima were at 488 and 507 nm, respectively). This fact indicates that the spectral properties of the fluorescent protein did not change when it was fused with α-CbTx. The yield of recombinant eGFP-α-CbTx was ∼10 mg per 1 L of bacterial culture; the purity of the final product was not less than 95%.
Modification of α-CbTx and Aze with p-HOBDI-BF 2 To obtain p-HOBDI-BF 2 -labeled toxins, they were reacted with p-HOBDI-BF 2 -OSu ( Figure 2) using conditions elaborated earlier for α-CbTx modification by activated esters of photoactivatable labels. The modified α-CbTx was purified using high-performance ion-exchange and reversed-phase chromatography as described earlier (Utkin et al., 1998), and similarly one predominant derivative (HOBDI-BF 2 -α-CbTx) was obtained. This derivative was analyzed by high-resolution mass spectrometry, which showed that it possessed a molecular mass of 8,112.76 Da. This value was 19.92 Da lower than the theoretical mass; the difference is explained by the loss of HF from fluorinecontaining label under mass spectroscopy conditions, this phenomenon was earlier described for the BODIPY label (Qi et al., 2015). Peptide mass fingerprinting using trypsin digestion revealed the labeled fragment 13-33 (data not shown), which indicates that α-CbTx derivative contains the label at Lys23. A similar technique was applied to introduce the p-HOBDI-BF 2 label into the molecule of amidated Aze (see structure in Figure 3A). The Aze derivatives were purified by reversed-phase HPLC ( Figure 3A) and analyzed by mass spectrometry. It was found that fraction 2 with the molecular mass of 2,539 Da represents unmodified toxin. Fraction 3 contains a di-labeled derivative (3,132 Da) contaminated with unmodified toxin. Fractions 4-6 contain different mono-modified products (HOBDI-BF 2 -Aze) with molecular masses of 2,836 Da, fraction 6 being contaminated with a di-labeled product. Interestingly, the mass corresponding to the loss of HF was also observed. As there are 3 amino groups (N-terminal and ε-amino groups of Lys6 and Lys20) in Aze, our results show that three corresponding mono-labeled derivatives were obtained.
According to chromato-mass spectrometry analysis, among all these fractions only peak 4 was an individual compound (data not shown). To localize the fluorophore in this mono-derivative, it was digested with two proteases, and the obtained fragments were analyzed by mass spectrometry. The digestion of this derivative with chymotrypsin resulted in a fragment with molecular mass of 2,420 Da containing the fluorescent label ( Figure 3B). This product corresponds to the C-terminal fragment 4-21 of HOBDI-BF 2 -Aze, which indicates that the N-terminal amino group is not modified. To find out which of the two lysine residues (Lys6 or Lys20) is modified, the derivative was digested with Lys-C endoproteinase, and the obtained products were analyzed by mass spectrometry. A peptide with molecular mass of 2,740 Da was found ( Figure 3C). It corresponds to the N-terminal fragment 1-20 containing the fluorescent label, suggesting that the fluorophore is attached to Lys6 (see structure in Figure 3A).

Competitive Radioligand Assay
We evaluated the affinity of the prepared fluorescent α-CbTx derivatives to the muscle-type nAChR from T. californica ray electric organ and neuronal human α7 nAChR expressed in GH 4 C 1 cells, as well as to AChBP from L. stagnalis, which is a homolog of the ligand-binding domain of all nAChRs. This was done by measuring their ability to compete with 125 I-α-BgTx used as a radioligand. When testing fluorescent α-CbTx (eGFP-α-CbTx and HOBDI-BF 2 -α-CbTx), α-CbTx and eGFP were also used in parallel experiments. The respective inhibition curves on muscle-type and α7 nAChRs, and AChBP are shown in Figures 4A-C and corresponding IC 50 values and slopes (Hill coefficients) are presented in Table 2.
Frontiers in Molecular Biosciences | www.frontiersin.org November 2021 | Volume 8 | Article 753283 when the N-terminal His-tag was removed from eGFP, this inhibitory activity disappeared (data not shown).
In order to figure out the reason for such a noticeable drop in affinity of eGFP-α-CbTx, α-CbTx was cleaved from the chimeric fusion protein eGFP-EK-α-CbTx and purified by reversed-phase HPLC. The affinity of purified recombinant α-CbTx towards muscle-type nAChR was markedly lower than that of the native α-CbTx (IC 50 value was >10 nM vs 0.57 ± 0.03 nM, respectively). This result suggests that recombinant α-CbTx in the chimeric eGFP-α-CbTx protein does not fold properly.
A competitive radioligand analysis was also used to evaluate the affinity of the obtained fluorescent Aze derivatives from fractions 3-6 ( Figure 3A) towards the muscle-type T. californica nAChR. The inhibition curves and the corresponding IC 50 values are shown in Figure 4D and Table 2. The affinity of the derivatives from fractions 4-6 decreased after modification by more than an order of magnitude (IC 50 of 170 ± 50 nM compared to IC 50 7.3 ± 0.2 nM for unmodified Aze). Fraction 3 showed a markedly higher affinity (IC 50 23 ± 2 nM), most likely due to the presence of an unmodified toxin detected by mass spectrometric analysis (see above). The small differences in affinity of mono-labeled derivatives from fractions 4-6 probably reflect the similar importance of modified amino groups for binding to the receptor.

Fluorescent Labeling of nAChRs in Cell Cultures and Tissue Slices
Fluorescence microscopy of Neuro-2a cells transfected with human α7 nAChR, and rat tongue preparation containing muscle fiber end plates with muscle α1β1εδ nAChR, revealed a difference in applicability of the HOBDI-BF 2 and eGFP derivatives of α-CbTx. Control commercial AF555-α-BgTx (50 nM) shows bright staining of Neuro-2a cells and rat tongue end plates ( Figure 6A). Both eGFP-α-CbTx chimeric protein (500 nM) and HOBDI-BF 2 -α-CbTx (100 nM) bind to Neuro-2a cells transfected with human α7 nAChR. This staining is abolished by 1 h preincubation with a 30-50 molar excess of α-CbTx, confirming the specificity of binding ( Figure 6A). eGFP alone does not stain cells ( Figure 6A), thus confirming that the staining could not be attributed to a non-specific uptake of the fluorescent protein.
eGFP-α-CbTx stained rat tongue end plates brightly and specifically ( Figure 6B). Surprisingly, we failed to achieve end-FIGURE 4 | Inhibition of the initial rate for 125 I-α-BgTx binding to (A) T. californica nAChR, (B) human α7 nAChR and (C) L. stagnalis AChBP with α-CbTx and its fluorescent analogs (eGFP-α-CbTx, HOBDI-BF 2 -α-CbTx) as well as eGFP. (D) Inhibition of initial rate for 125 I-α-BgTx binding to T. californica nAChR with Aze and its p-HOBDI-BF 2 derivatives from fractions 3-6 (see Figure 3A). Each point is a mean ± S.E.M. value of three measurements for each concentration. The curves were calculated from the means ± S.E.M. using OriginPro 2015, and the respective IC 50 values in nM are presented in Table 2. TABLE 2 | Affinity of the fluorescent toxins tested in competition with 125 I-α-BgTx for binding to muscle-type nAChR from T. californica, human neuronal α7 nAChR, and AChBP from L. stagnalis calculated from the respective inhibition curves in Figure 4 as the IC 50 values and Hill coefficients (n H ) using OriginPro 2015. plate staining with HOBDI-BF 2 -α-CbTx. We also tried to stain end plates with HOBDI-BF 2 -Aze, however, no end-plate staining was detected in this case either ( Figure 6B). It is worth noting that rat tongue preparations are made using denaturating conditions (isopropyl alcohol). The absence of specific staining of rat tongue cross-sections by the HOBDI-BF 2 derivatives of neurotoxins could be explained by the intrinsic properties of this fluorophore preventing it from binding to partially denaturated muscle nAChR. On the other hand, the absence of staining by HOBDI-BF 2 -α-CbTx or HOBDI-BF 2 -Aze could be attributed to their fast wash-out kinetics, which might be due to the modification of lysine side chains crucial to the binding.
Next, we performed binding assays using eGFP-α-CbTx to measure the surface expression of α7 nAChR levels on Raji and RPMI 1788 cells. Figure 7B shows that eGFP-α-CbTx can stain both cell lines (red histograms), and this binding is specific because it decreased after pre-incubation with an excess of free α-CbTx (green histograms). However, again the calculated average values for Raji and RPMI 1788 cells labeled with fluorescent toxin (positive cells) were only ∼5 and 9%, respectively ( Figure 7C). The average values of specific binding of eGFP-α-CbTx for Raji, RPMI 1788, and THP-1 Mϕ cells were 41, 28, and 30%, respectively ( Figure 7D). Flow cytometry measurements showed that eGFP-α-CbTx specifically stains cell lines that express α7 nAChR ( Figure 7C). On the other hand, HOBDI-BF 2 -α-CbTx was found unsuitable for this purpose.

DISCUSSION
Studies of ligand-receptor interactions require adequate molecular instruments and methods. Radioligand analysis was the method of choice for such studies for a long time. However, this method has limitations related to safety and ecological aspects. Fluorescence-based methods became more convenient and popular due to the development of new approaches including super-resolution techniques such as Photoactivated Localization Microscopy (PALM) and Stochastic Optical Resolution Microscopy (STORM) and highly fluorescent labels (e.g., quantum dots) (see reviews (Baddeley and Bewersdorf, 2018;Himmelstoß and Hirsch, 2019)). Among the developed methodologies is the application of fluorescent proteins (GFP and analogs) for the labeling. Modern fluorescent proteins cover a wide range of excitation and emission wavelengths and possess high quantum yields (Chudakov et al., 2010). Recently, we proposed a novel approach for toxin labeling, which is based on the production of chimeric proteins consisting of two modules: a fluorescent protein and a polypeptide toxin specific for certain channels where the ligand part serves for selective channel recognition, and the fluorescent part is used for effective detection. Chimeric fusion proteins comprising GFP (or analogs) and the protein ligand under the study can be easily produced recombinantly in high quantities. Relying on this approach we produced a chimera (eGFP-OSK1) for potassium channel recognition (Kuzmenkov et al., 2016a). Further, this approach was also applied for K V 1.3-specific chimera production due to clinical importance of this isoform (Nekrasova et al., 2020).
The preparation of fluorescently labeled proteins and peptides by chemical modification is another frequently used method for which appropriate reagents have long been developed. N-Hydroxysuccinimide esters of respective fluorophores are most often used, resulting in modification of amino groups (the N-terminal α-amino group or ε-amino groups of lysine residues). This method allows introducing labels into different parts of the molecule, which is not always possible with fluorescent proteins. However, sometimes the introduction of large fluorophores into peptides or small proteins can lead to a loss of activity. Nevertheless, fluorescently labeled toxins were successfully used to study toxinreceptor interactions. We were among the first in application of fluorescently labeled TFTs for the investigation of nAChRs (Tsetlin et al., 1979(Tsetlin et al., , 1982Surin et al., 1983) and we continue to successfully use them for this purpose until now (Shelukhina et al., 2009;Pivovarov et al., 2020). The first studies showed the preferential modification of the lysine in the central loop of TFTs, which did not lead to a significant change in the affinity to the receptor, and allowed to characterize the toxin site involved in the interaction with nAChR.
When introducing bulk moieties into small peptides, a noticeable decrease in affinity, a change in specificity or an increase in non-specific interactions is more often observed, as we demonstrated earlier with the example of some photoactivated derivatives of α-conotoxins MI and GI (Kasheverov et al., 2001(Kasheverov et al., , 2006. However, fluorescent derivatives of some α-conotoxins were prepared, which became selective markers of different nAChR subtypes (Hone et al., 2010;Surin et al., 2012;Barloscio et al., 2017). For instance, using a fluorescent derivative of α-conotoxin MII, a selective α6β2 nAChR antagonist, α6-containing nAChRs were localized in the retinal tissue and found to be predominantly localized to the ganglion cell layer (Barloscio et al., 2017).
The introduction of a bulky fluorescent label can result in a decrease of affinity and change of selectivity of toxins. However cryo-EM structure of Torpedo acetylcholine receptor in complex with α-bungarotoxin (Rahman et al., 2020) reveals that both Nand C-termini of TFTs are less important for the interaction with the receptor and can tolerate the attachment of a sizeable substituent. Keeping this in mind we prepared a chimeric protein, in which eGFP was fused with α-CbTx N-terminus through a linker of 19 amino acid residues (Figure 1). In radioligand assays, the chimeric protein with a full-sized GFP FIGURE 6 | Fluorescence microscopy of Neuro-2a cells transfected with human α7 nAChR, and rat tongue preparation containing muscle fiber end plates with muscle α1β1εδ nAChR. (A) Control staining of Neuro-2a cells with 50 nM AF555-α-BgTx shows peripheral staining of the cells, which is reproduced by the eGFP-α-CbTx chimeric protein (500 nM). eGFP alone does not stain the cells. Complementary to that, specific eGFP-α-CbTx staining is abolished by 1 h preincubation with 15 µM α-CbTx. HOBDI-BF 2 -α-CbTx (100 nM) also stains Neuro-2a cells, and it is prevented by 1 h pre-treatment using 5 µM α-CbTx. Scale bar in the first panel, 50 μm. (B) AF555-α-BgTx (100 nM) specifically stains end plates in rat tongue cross-sections, and no staining is observed in cross-sections pre-treated with 5 µM α-CbTx. Neither HOBDI-BF 2 -α-CbTx, nor HOBDI-BF 2 -Aze produce end plate staining, whereas 500 nM eGFP-α-CbTx chimeric protein (12 h of incubation) stained cross-sections brightly. End-plate staining by eGFP-α-CbTx chimeric protein is prevented by 1 h pre-treatment with 15 µM α-CbTx. Scale bar in the first panel, 50 μm.
Frontiers in Molecular Biosciences | www.frontiersin.org November 2021 | Volume 8 | Article 753283 (eGFP-α-CbTx) showed a significantly diminished affinity towards all tested targets as compared to α-CbTx. The decrease in affinity for muscle-type T. californica, human α7 nAChR, and L. stagnalis AChBP was ∼272, 169 and 9.5 times (Figures 4A-C; Table 2). To find out the reasons for this discrepancy, the toxin was split from eGFP and purified. The obtained recombinant α-CbTx was substantially less active than the native toxin, leading us to the assumption that an essential part of the recombinant toxin was incorrectly folded.
As mentioned in the Introduction, the synthetic analog of GFP chromophore p-HOBDI-BF 2 was prepared earlier (Baranov et al., 2012). This label is similar in many of its spectral characteristics to the well-known dye fluorescein. It has a similar spectrum, close extinction coefficient, and quantum yield of fluorescence. However, compared to fluorescein, HOBDI-BF 2 is minute, its molecular mass is 4 times smaller. It is also assumed as more biomimetic since it is based on the GFP chromophore. Thus, the use of this label should affect the properties and behavior of the stained biological objects to a significantly lower degree. To utilize this chromophore for specific amino group labeling in proteins, an activated N-hydroxysuccinimide ester was synthesized. We introduced p-HOBDI-BF 2 into α-CbTx, and one predominant fluorescent derivative HOBDI-BF 2 -α-CbTx was obtained. Compared to native α-CbTx, this derivative showed a reduced affinity to muscle-type T. californica, human α7 nAChR, and L. stagnalis AChBP by ∼6.8, 5.0 and 2.4 times, respectively (Figures 4A-C; Table 2), still retaining a high affinity to all tested targets. Since α-CbTx binds to both muscletype and some neuronal nAChRs, we also prepared fluorescent derivatives of Aze, a selective antagonist of muscle-type nAChR. p-HOBDI-BF 2 introduction into Aze led to the formation of multiple reaction products (mono-and di-derivatives with possible label localization at the α-amino group of Asp1 and ε-amino groups of Lys6 or Lys20; Figure 3A). To identify the exact localization of the label for the best purified derivative (fraction 4), a chromato-mass spectrometru analysis of its chymotrypsin ( Figure 3B) and Lys-C fragments ( Figure 3C) was performed. Interestingly, the introduction of a p-HOBDI-BF 2 chromophore into a small peptide Aze did not dramatically decrease its affinity to muscle-type nAChR ( Figure 4D).
The evaluation of the binding of two fluorescent toxins (eGFP-α-CbTx and HOBDI-BF 2 -α-CbTx) to AChBP using MST ( Figure 5) showed that the affinity of the photoprobes lies in the same concentration range, i.e. 10-200 nM. This result suggests that both methods of obtaining fluorescent products (recombinant and chemical modification) used by us lead to sufficiently effective tools for detecting the respective receptor targets.
Optical microscopy is one of the major fields for the application of fluorescent derivatives of neurotoxins. Both HOBDI-BF 2 -α-CbTx and eGFP-α-CbTx were applicable for the detection of nAChR in live cell imaging ( Figure 6A). eGFP-α-CbTx is able to detect muscle nAChR in a tissue preparation and represents a cheap analog of the chemically modified α-BgTx conventionally utilized for this task ( Figure 6B). However, neither HOBDI-BF 2 -α-CbTx nor HOBDI-BF 2 -Aze were able to stain histological cross-sections of rat tongue containing muscle nAChR. Their inability to bind to nAChRs that are present in rat tongue preparations could be explained by partial denaturation of the receptor due to isopropyl alcohol treatment. As a result of such treatment some fluorescent toxins retain the ability to bind to muscle nAChR (AF555-α-BgTx, eGFP-α-CbTx), but others lose their affinity. These features of new fluorescent derivatives should be taken into account in experimental design.
In flow cytometry experiments (Figure 7), it was revealed that a very small proportion (<10%) of the three studied cell types (Raji, RPMI 1788, and THP-1 Mϕ) expressing α7 nAChR (Siniavin et al., 2020), is specifically stained by eGFP-α-CbTx. The use of HOBDI-BF 2 -α-CbTx produced results at the level of autofluorescence ( Figure 7A). This is in particular contrast to the commercially available AF647-α-BgTx ( Figures 7A,C), which stained ∼60% of cells. Thus, there are applications where our novel derivatives are efficient, but there are numerous cases where traditional readioactive and fluorescent labels retain their role.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The animal study was reviewed and approved by Institutional Policy on the Use of Laboratory Animals of the Shemyakin-Ovchinnikov Institute of Bioorganic Chemistry Russian Academy of Sciences (Protocol Number 327/2021).

AUTHOR CONTRIBUTIONS
IK, AK, DK, AV, and YU designed research; IK, AK, DK, AV, and YU analyzed data; IK, AK, DK, AV, and YU performed research; IK, AK, DK, AV, and YU wrote the paper; IC performed gene engineering and chimera production; IS, DK, performed cyto and histochemistry; IK synthesized radioligand and performed radioligand analysis; EB obtained thermophoresis spectra; II synthesized peptides and localized the fluorescent label, AS performed flow-cytometry, RZ measured molecular masses, MB performed fluorescent label synthesis, VT, AV, and YU supervised research.

FUNDING
This study was funded by RFBR, project number 20-54-00033. Fluorescent chimera design and production were supported by RFBR, project number 20-34-70031.