α-CbTx was purified from N. kaouthia venom as described (Osipov et al., 2008). Aze was synthesized as in (Shelukhina et al., 2018). Radiolabeled ¹²⁵I-α-BgTx with specific radioactivity of 500 Ci/mmol was prepared as in (Kryukova et al., 2018). Alexa Fluor 555 and 647 conjugates of α-BgTx (AF555–α-BgTx and AF647–α-BgTx) were purchased from Thermo Fisher Scientific.

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-BF₂ 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).

Murine neuroblastoma Neuro-2a cells were obtained from ATCC-LGC (France) and maintained as described previously (Shelukhina et al., 2017). THP-1 (ATCC TIB-202), Raji (ATCC CCL-86), and RPMI 1788 (ATCC CCL-156) cells were maintained in RPMI 1640 medium, supplemented with 10% fetal bovine serum, 1× GlutaMAX and 1× penicillin-streptomycin solution (all from Gibco, Thermo Fisher Scientific) at 37°C with 5% CO₂.

At the first step the pET-28a-eGFP vector was obtained by cloning the eGFP gene alone, which was amplified with eGFP-f and eGFP-r 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.

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–EK-α-CbTx preparations was measured by absorption spectroscopy using ε (489 nm) = 55,000 M⁻¹cm⁻¹.

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 reversed-phase 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.

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-BF₂ 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.

Kieselgel 60 (Merck, Germany) was used for column chromatography. Thin-layer chromatography was performed on silica gel 60 F₂₅₄ glass-backed plates (Merck). Visualization was performed using UV light (254 or 312 nm) and staining with KMnO₄. NMR spectra were recorded on a 700 MHz Bruker Avance III instrument at 303 K. Chemical shifts are reported relative to residual peaks of DMSO-d₆ (2.51 ppm for ¹H and 39.5 ppm for ¹³C). Melting points were measured on a Stuart SMP30 apparatus (Cole-Parmer) without correction. IUPAC compound names were generated using ChemDraw software (PerkinElmer).

1.0 mmol of (Z)-4-(4-(2-(difluoroboryl)-4-hydroxybenzylidene)-2-methyl-5-oxo-4,5-dihydro-1H-imidazol-1-yl)butanoic acid (Frizler et al., 2013) was dissolved in dry THF (30 ml), and N,N,N′,N′-tetramethyl-O-(N-succinimidyl)uronium tetrafluoroborate (330 mg, 1.1 mmol) and triethylamine (120 mg, 1.2 mmol) were added. The mixture was stirred at room temperature for 1 h. Hydrofluoric acid (water solution, 48%, 0.5 ml) was added and the mixture was stirred for additional 10 min. The mixture was dissolved by EtOAc (150 ml), washed by brine (3 × 30 ml), dried over Na₂SO₄, and evaporated. The product (N-hydroxysuccinimide ester of p-HOBDI-BF₂, p-HOBDI-BF₂-OSu) was purified by flash chromatography (using EtOAc:THF = 2:1 mixture as an eluent): yellow solid, 370 mg (85%), m. p. = 163–165°С.

¹H NMR (DMSO-d₆) δ 10.19 (s, 1H), 7.55 (s, 1H), 7.49 (d, J = 8.1Hz, 1H), 7.00 (d, J = 2.5Hz, 1H), 6.74 (dd, J₁ = 8.2Hz, J₂ = 2.5Hz, 1H), 3.80 (t, J = 7.27Hz, 2H), 2.85 (t, J = 7.48Hz, 2H), 2.82 (br.s, 4H), 2.74 (s, 3H), 1.99 (p, J = 7.48Hz, 2H).

¹³C NMR (DMSO-d₆) δ 12.65 (CH₃), 22.90 (CH₂), 25.40 (2×CH₂), 27.52, 39.88, 115.12 (CH), 118.48 (CH), 124.00, 124.87, 129.25 (CH), 134.27 (CH), 161.41, 162.93, 164.18, 168.51, 170.11.

High resolution mass spectrometry (ESI) [M + H]⁺ m/z: 434.1341 found; 434.1335 calculated for C₁₉H₁₉BF₂N₃O₆.

α-CbTx was modified with p-HOBDI-BF₂-OSu and the derivatives were purified exactly under conditions used earlier for the preparation of α-CbTx mono-labeled derivatives (Utkin et al., 1998). To obtain Aze derivatives, 5.3 mg of Aze were dissolved in 1 ml of 0.2 M sodium phosphate buffer (pH 8.2) and 1.1 mg of p-HOBDI-BF₂-OSu in 100 µl of dioxane was added to the toxin solution. The mixture was incubated overnight at room temperature and desalted on a Sephadex G-10 column (1 × 90 cm) equilibrated in 0.1 M acetic acid. After freeze-drying the mixture was separated by reversed-phase HPLC on a Jupiter C₁₈ column (10 × 250 mm; Phenomenex) in a linear gradient of acetonitrile (15–30% in 75 min) in the presence of 0.1% trifluoroacetic acid, at a flow rate of 2 ml/min. After freeze-drying, the obtained derivatives (HOBDI-BF₂–Aze) were used for further studies.

Purified HOBDI-BF₂–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.

In competition experiments with ¹²⁵I-α-BgTx (500 Ci/mmol), all tested compounds (α-CbTx, HOBDI-BF₂–α-CbTx, eGFP–α-CbTx, eGFP, Aze, and HOBDI-BF₂–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 GH₄C₁ 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 ¹²⁵I-α-BgTx was added to AChBP, GH₄C₁ 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 GH₄C₁ cells or membranes), unbound radioactivity being removed from the filters by washout (3 × 3 ml) with the binding buffer. Non-specific binding was determined in all cases using 2.5 h pre-incubation 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: % r e s p o n s e = 100 / { 1 + ( [ t o x i n ] / I C 50 ) n H } , where n _H is the slope factor (Hill coefficient). Data are presented as means ± standard errors of means (S.E.M.).

For the microscale thermophoresis (MST) experiments, both the fluorescent ligands (HOBDI-BF₂–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-BF₂–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.

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 carbonate-buffered 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).

For differentiation of THP-1 into macrophages (THP-1 Mϕ), cells were seeded in 24-well plates and treated with 100 nM phorbol 12-myristate 13-acetate (Sigma-Aldrich) for 48 h. Сells were stained as previously described (Siniavin et al., 2020). Briefly, THP-1 Mϕ, Raji, or RPMI 1788 cells were incubated with 100 nM HOBDI-BF₂–α-CbTx, eGFP–α-CbTx, eGFP, or AF647–α-BgTx (1 h, at 4°C). Cells were then washed two times with PBS and assayed using BD FACSCalibur or MACSQuant Analyzer 10 flow cytometer (Miltenyi Biotec, Germany) and analyzed using FlowJo software 10.0.8. To determine non-specific binding, cells were pre-incubated with 10 µM α-CbTx before fluorescent toxin staining.

A model of eGFP–α-CbTx was constructed using UCSF Chimera 1.10.1 software (Pettersen et al., 2004) for Modeller 9.14 (Webb and Sali, 2016). We utilized eGFP (PDB ID: 2Y0G) and α-CbTx (PDB ID: 1CTX) spatial structures as templates. The N-terminal site of chimeric protein and the (G₄S)₃-linker between the eGFP and α-CbTx modules were represented in a disordered conformation. A model of eGFP–α-CbTx was visualized by the PyMOL Molecular Graphics System, Version 1.7.4 Schrödinger, LLC.

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%.

To obtain p-HOBDI-BF₂-labeled toxins, they were reacted with p-HOBDI-BF₂-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-BF₂–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 fluorine-containing 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-BF₂ 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-BF₂–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-BF₂–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).