INTRODUCTION

One of the main applications of venom in snakes is hunting. The most effective way to immobilize prey is to disrupt the function of the nervous system. This strategy is used by snakes of the family Elapidae, the venom of which contains neurotoxins blocking neuromuscular transmission. Snakes of this family have a well-developed venom-injecting apparatus including front fangs and hence are referred to as front-fanged snakes (proteroglyphous snakes in the English classification). In the Russian classification, snakes of the family Viperidae (solenoglyphous snakes in the English classification) with mainly hemotoxic venoms are also referred to as front-fanged snakes. The bites of these snakes are very dangerous to humans. It should be noted that there is a large group of snakes classified into different families, which have venom fangs located in the back of the jaw, the so called rear-fanged snakes (opistoglyphous snakes in the English classification). Most of them are not dangerous to humans, and some species are kept as pets. Thus, a quite popular and frequent species for terrarium keepers is the Malagasy cat-eyed snake Madagascarophis colubrinus.

M. colubrinus is a species of snakes of the family Lamprophiidae, the subfamily Pseudoxyrhophiidae, living mainly on Madagascar [1]. These snakes are found in a variety of habitats throughout the island of Madagascar, from mountainous areas to tropical forests. They are mainly nocturnal terrestrial snakes that can climb and swim well. They feed on a variety of prey, including chameleons and rodents. Their venom is weak, often not potent enough to kill the chosen prey, so they also use compression whenever required. For humans, the bite of this snake is mildly venomous [2]. Envenomation can cause acute local effects, including pain, swelling, blistering, and tissue necrosis [3].

Previously, a number of neurotoxins have been identified in the venoms of rear-fanged snakes. For example, neurotoxins from the family of three-finger toxins were isolated from the venoms of snakes of the genus Boiga, the family Colubridae [4]. Neurotoxins of the same family were identified in the venom of the green vine snake Oxybelis fulgidus [5] and the coral mimic snake Rhinobothryum bovallii [6]; both these species also belong to the family Colubridae. It should be noted that, in spite of transcriptomic [7] and proteomic [8] analyses of the venoms of rear-fanged snakes of the family Lamprophiidae, there are no data on identification and isolation of neurotoxins from their venoms.

As regards the neurotoxins of Viperidae snakes, we have previously shown that phospholipases A2 from the venoms of these snakes are able to inhibit nicotinic acetylcholine receptors (nAChR) [9, 10]. We have isolated several other different peptide and protein inhibitors of nAChR from venoms of the snakes of this family [11–13]. These data indicate that the venoms of Viperidae snakes are promising as a source of neuroactive compounds.

The aim of the present work was to study the neurotoxic activity of venoms of some snakes from the family Viperidae and the Malagasy cat-eyed snake M. colubrinus and to isolate the toxin with neurotoxic potential from the venom of M. colubrinus. Since neurotoxicity is most pronounced when neuromuscular transmission is blocked, we have used the inhibition of the function of nAChR, which is the key element of neuromuscular transmission, as a test.

RESULTS AND DISCUSSION

Interaction between snake venoms and nicotinic acetylcholine receptor. In the present work, the activity of venoms and their components was determined by the competition with α-bungarotoxin (α-Bgt) for the binding to nAChR. This toxin is a highly effective antagonist of neuronal nAChRs of types α7 and α9, as well as muscle-type nAChRs, and is still widely used as a marker of these receptors [14]. The target of α-Bgt in our work is muscle-type nAChRs. Previously, we have successfully used fluorescently labeled α-Bgt to study toxin–receptor interactions [13, 15]. In the present work, we have used a α-Bgt derivative fluorescently labeled with the ALEXA 488 dye, hereafter referred to as ALEXA488-Bgt.

The competitive fluorescence analysis using nAChR of the electric organ of Torpedo californica was performed as described previously [13]. Along with the venom of the Malagasy cat-eyed snake M. colubrinus, we also tested the venoms of the mamushi Glodyus blomhoffi, the rock mamushi G. saxatilis, Orlov’s viper Vipera orlovi, the habu Protobothrops flavoviridis and the South American bushmaster Lachesis muta. As follows from the data obtained (Fig. 1), the venom of the Malagasy cat-eyed snake M. colubrinus, as well as the venoms of the mammushi G. blomhoffi and the rock mammushi G. saxatilis, demonstrated the remarkable ability to inhibit the binding of ALEXA488-Bgt to the nAChR of T. californica. The venom of the Malagasy cat-eyed snake M. colubrinus turned out to be one of the most active against muscle-type nAChRs, displacing 75% of ALEXA488-Bgt at a concentration of 2 mg/mL. The venom of the mammushi G. blomhoffi showed high activity, inhibiting 65% of ALEXA488-Bgt binding at a concentration of 2 mg/mL. The venom of the rock mammushi G. saxatilis showed even greater efficiency of binding to the muscle-type receptor, completely displacing ALEXA488-Bgt at a concentration of 2 mg/mL.

Fig. 1.
figure 1

The competition of snake venoms with ALEXA488-Bgt for the binding to the nAChR of T. californica.

Full size image

Based on the results of fluorescence competition assay, it can be concluded that the venoms contain components competing with ALEXA488-Bgt for the binding to nAChR; however, it is impossible to determine whether the active components of venoms under study are acetylcholine receptor agonists or antagonists. The venoms showing the most effective inhibition of ALEXA488-Bgt binding to the nAChR of Torpedo were chosen for further work. Calcium imaging was used to determine whether these venoms affect the functional response of the receptor. The study was performed in human embryonic kidney (HEK293) and mouse neuroblastoma (Neuro2a) cell lines transfected with the plasmids encoding human muscle nAChR subunits at a ratio of 2α1 : 1β1 : 1δ : 1ε, as well as the Case12 calcium sensor. When the receptor is activated, calcium enters the cell through an open channel and causes an increase in the sensor fluorescence. The sensor allows direct estimation of the changes in Ca2+ concentration within a range from nano- to micromoles with a high signal-to-noise ratio. The rapid and reversible binding of Case12 [16] to calcium ions allows the sensor to be used for monitoring the level of calcium oscillations in the cell. In response to an increase in Ca2+ concentration, there is an increase in fluorescence intensity of the sensor up to 12-fold.

The cells that responded to the addition of 40 μM of acetylcholine by an increase in fluorescence emission, which indicated the presence of an expressed receptor on their surface, were chosen for the study. Then the cells were preincubated with the venoms for 5 min, followed by the addition of acetylcholine at the same concentration and determination of the response to acetylcholine. The results were recorded for each individual cell. It was established that preincubation of cells of the lines under study with the venom of the rock mammushi G. saxatilis inhibited cell response to acetylcholine by 73 and 85% for HEK293 and Neuro2a cells, respectively (Fig. 2). In case of preincubation of cells with the venom of the Malagasy cat-eyed snake M. columbrinus, the response to acetylcholine was inhibited by 75 and 100% for HEK293 and Neuro2a cells, respectively (Fig. 2).

Fig. 2.
figure 2

The changes in fluorescence intensity corresponding to the changes in the intracellular concentration of calcium ions in HEK293 and Neuro2a cells. Different colors indicate the responses from individual cells. (a, b, e, f) The responses to 5-s incubation with 40 µM acetylcholine; (c, d, g, h) the responses to 5-s incubation with 40 µM acetylcholine after 5-min preincubation with venoms of the snakes G. saxatilis (c, d) and M. colubrinus (g, h).

Full size image

Thus, the rock mammushi and cat-eyed snake venoms inhibited the functional response of the human muscle-type nAChR, exhibiting the properties of competitive antagonists. The results of further studies on the rock mammushi venom will be published elsewhere. In the present work, the cat-eyed snake venom was further separated by liquid chromatography to isolate the active component.

Isolation of the nAChR peptide inhibitor from cat-eyed snake venom and determination of its amino acid sequence. To isolate the inhibitor, the crude venom of M. colubrinus was separated by three stages of liquid chromatography. At the first stage, gel filtration on a Superdex 75 column was used (Fig. 3a), then the ability of the obtained fractions to compete with α-Bgt for binding to Torpedo nAChR was determined. Fraction II proved to be most active and was further separated by reversed-phase high-performance liquid chromatography (Fig. 3b); the fractions were also analyzed for the ability to compete with α-Bgt for the binding to the Torpedo nAChR. Fraction 18 showed the greatest activity and was additionally purified by reversed-phase high-performance liquid chromatography using a smoother acetonitrile concentration gradient (Fig. 4).

Fig. 3.
figure 3

The isolation of active compound from the venom of M. colubrinus: (a) gel filtration on a Superdex 75 column (10 × 300 mm) equilibrated with 0.1 M ammonium acetate (pH 6.2) at a flow rate of 0.5 mL/min; (b) separation of fraction II by reversed-phase chromatography on a Jupiter C18 column (10 × 250 mm) in a 10–55% acetonitrile concentration gradient for 90 min at a flow rate of 2 mL/min. The horizontal lines under the peaks indicate active fractions.

Full size image
Fig. 4.
figure 4

The separation of fraction 18 (Fig. 3b) by reversed-phase chromatography on a Jupiter C18 column (4.6 × 250 mm) in an acetonitrile concentration gradient of 20–30% for 60 min at a flow rate of 1 mL/min. The horizontal line indicates the active fraction.

Full size image

Fraction 5 (Fig. 4) was able to inhibit the binding of α-Bgt with the Torpedo nAChR and underwent further analysis. According to mass spectrometry data, the isolated product had a molecular mass of 2786.3 Da. This mass corresponds to a peptide of ~25 AAR. Taking this fact into account, the amino acid sequence of the peptide was determined by the automated Edman degradation with a PPSQ-33A protein and peptide sequencer (Shimadzu Corp., Japan). The identified amino acid sequence of the peptide named macoluxin (from Madagascarophis colubrinus toxin) is shown in Fig. 5a. The calculated molecular weight of macoluxin was 2786.2 Da; within the error, it coincides with the experimentally determined value. Next, a search was made for homologous amino acid sequences in non-redundant protein sequence databases including GenBank, PDB, SwissProt, PIR and PRF, using the BLAST algorithm [17] (https://blast.ncbi.nlm.nih.gov/Blast.cgi).

Fig. 5.
figure 5

(a) The amino acid sequence of macoluxin; (b) the alignment of the macoluxin amino acid sequence with homologous fragments of the known proteins: QSI84039.1, the metalloproteinase of Calliophis bivirgatus; ABK63559.1, the metalloproteinase of Demansia vestigiata; D6PXE8.1, the zinc metalloproteinase-disintegrin-like ATPase B of Naja atra; KAG5858171.1, the metalloproteinase of Bothrops jararaca; Q8QG89.1, the metalloproteinase BITM02A of Bothrops insularis; Q2QA02.1, the metalloproteinase of Crotalus durissus durissus.

Full size image

The search results have shown that macoluxin is homologous to a fragment of the amino acid sequence of metalloproteinases of the venoms of snakes from different families (Fig. 5b). At the same time, the degree of similarity is very high and reaches 91%. Interestingly, in the spatial structure of metalloproteinases, this fragment forms an α-helix located on the surface of the protein globule (e.g., the 3K7N_A structure [18]). This probably explains the cleavage of a fragment of this size. It is interesting to note that peptides Pm1 and Pm2 capable of inhibiting nAChR have been previously identified in the propeptide domain of the metalloproteinase of the olive sand snake Psammophis mossambicus when analyzing the gene evolution of these multifunctional enzymes [19]. During biosynthesis, the propeptide domain is cleaved from the precursor protein with the formation of a mature metalloproteinase with a proteolytic activity. The enzymes involved in snake venom protein processing have been little studied; in particular, it is unknown what proteases cleave the propeptide domain. Metalloproteinases are widespread in snake venoms, and their levels in some venoms are very high, but the data on identification of propeptide domain fragments in venoms are very limited. Such fragments have been found, in particular, by the proteomic analysis of the venom of Bothrops jararaca [20]. As regards peptides Pm1 and Pm2, there are no data on the presence of these compounds in the venom of the snake P. mossambicus. This snake is rear-fanged similar to M. colubrinus but belongs to the family Psammophiidae. Since macoluxin and peptides Pm1 and Pm2 have been identified in different domains of metalloproteinases, there is no homology between them. Nevertheless, the detection of these neurotoxic peptides indicates that the large proteins of venoms can be precursors of peptides with activities different from that of the original toxin. In particular, metalloproteinases (macoluxin seems to be a fragment of one of them) are a group of multidomain proteins exhibiting several types of biological activity such as the ability to cause hemorrhage, proteolytic degradation of fibrinogen and fibrin, induction of apoptosis, and inhibition of platelet aggregation. No data on the neurotoxic activity of metalloproteinases could be found.

As it has been mentioned in the introduction, three-finger toxins (nAChR blockers) were previously identified in the venoms of rear-fanged snakes [4–6]. These toxins are proteins with molecular masses of ~10 kDa containing five disulfide bonds, while macoluxin has no disulfide bonds. No close structural analogs of macoluxin have been found among the toxins acting on nAChRs. It should be noted that besides Pm1 and Pm2, previously we have isolated, the peptides inhibiting nAChR and having no disulfide bonds from the venoms of the black-headed Burmese viper Azemiops feae (the peptide azemiopsin) [11] and the puff adder Bitis arietans (baptides) [12]. However, these snakes, in contrast to M. columbrinus, belong to the family Viperidae. Another example of peptides, which do not contain disulfide bonds and inhibit nAChR, is conorfamides from the venom of the Mexican sea snail Conus austini [21]. All these peptides, with the exception of baptides, are characterized by the high content of basic amino acids, but there is no homology between them.

Interaction between macoluxin and nAChR. For a more detailed study of biological activity, the peptide macoluxin was obtained by solid-phase peptide synthesis. The peptide purity and structure were confirmed by analytical reversed-phase chromatography and mass spectrometry, respectively.

The efficiency of macoluxin interaction with nAChRs was determined by its competition with radioactive α-bungarotoxin (125I-αBgt) for binding to the membranes of the electrical organ of the ray T. californica containing muscle-type receptors (α12β1γδ) and with the GH4C1 cells expressing human neuronal α7 nAChR. The peptide was shown to inhibit the binding of 125I-αBgt to the membrane of Torpedo with IC50 = 46.8 ± 3.9 μM (Fig. 6). Macoluxin at a concentration of 50 μM did not inhibit the binding of 125I-αBgt to α7 nAChR; at a concentration of 100 μM, there was only 10% inhibition. It is interesting that, in contrast to macoluxin, the previously mentioned peptides Pm1 and Pm2 inhibited human α7 nAChR, showing the IC50 values of ~ 12 μM [19].

Fig. 6.
figure 6

The inhibition of the binding of radioactive α-Bgt to the nAChR of T. californica by macoluxin (IC50 = 46.8 ± 3.9 μM).

Full size image

The fact that macoluxin is a functional inhibitor of nAChR was verified by electrophysiological experiments with muscle-type nAChR heterologously expressed in African clawed frog oocytes. The peptide per se did not induce ionic currents. However, macoluxin inhibited the current induced by acetylcholine (Fig. 7). Moreover, its inhibitory effect completely disappeared after 5-min standard washing. Consequently, macoluxin is a reversible muscle-type nAChR antagonist.

Fig 7.
figure 7

Macoluxin inhibition of acetylcholine (ACh)-induced currents in muscle-type nAChR. The oocyte response to the addition of 30 μM acetylcholine was recorded, followed by washing (at least 5 times by 100 μL), incubation with macoluxin (5 min), and addition of macoluxin mixed with 30 μM acetylcholine.

Full size image

Thus, the peptide macoluxin isolated from the venom of the cat-eyed snake M. colubrinus is able to reversibly inhibit muscle-type nAChR. Macoluxin has a high degree of similarity with the fragment of snake venom metalloproteinase and seems to be formed as a result of proteolysis of this enzyme.

EXPERIMENTAL

Materials. The membranes of Torpedo californica were kindly provided by Professor F. Huho (Freie Universität Berlin, Germany). ALEXA 488-labeled α-Bgt (ALEXA488-Bgt) was purchased from Thermo Fisher Scientific (Waltham, United States). All reagents used in this work were of the “analytically pure” or higher grade.

Venom extraction from M. colubrinus. The venom was obtained from M. colubrinus specimens kept in the serpentarium of the Shemyakin–Ovchinnikov Institute of Bioorganic Chemistry, Russian Academy of Sciences. Venom samples were dried over anhydrous CaCl2 and stored at –20°С.

Peptide isolation. The venom was dissolved in 0.1 M ammonium acetate (pH 6.2) and applied to a Superdex 75 column (10 × 300 mm; Cytiva, United States) equilibrated with the same buffer. Elution was performed at a flow rate of 0.5 mL/min. The optical density of the eluent was recorded at 226 nm with an Uvocord SII UV flow detector (LKB, Sweden). Fraction II (Fig. 3a) was lyophilized and further separated by reversed-phase chromatography on a Jupiter C18 column (10 × 250 mm; Phenomenex, United States) in an acetonitrile gradient of 0–55% for 90 min in the presence of 0.1% trifluoroacetic acid at a flow rate of 2 mL/min (Fig. 3b). Fraction 18 (Fig. 3b) was lyophilized and further purified by reversed-phase chromatography on a Jupiter C18 column (4.6 × 250 mm; Phenomenex, United States) in an acetonitrile gradient of 20–30% for 60 min in the presence of 0.1% trifluoroacetic acid at a flow rate of 1 mL/min (Fig. 4).

Amino acid sequencing. The primary structure of the peptide was determined by Edman degradation with a PPSQ-33A automated protein and peptide sequencer (Shimadzu Corp., Japan) according to the manufacturer’s protocol.

Peptide synthesis. The peptide macoluxin was synthesized as described [22].

Biological activity of macoluxin. Competitive fluorescence analysis. All experiments were performed in a 96-well ultra-low attachment round-bottom plate (MICROTEST TM U-Bottom, Thomas Scientific, United States). The binding buffer contained: 137 mM NaCl, 2.7 mM KCl, 1.46 mM KH2PO4, 10 mM Na2HPO4, 0.1% Tween-20, pH 7.4. The tested membrane solution at a concentration of 0.125 µg/mL of protein (0.31 nM of toxin-binding sites) was added to the plate wells and incubated for one hour in 100 µL of the binding buffer with the G. blomhoffi, G. saxatilis, V. orlovi, M. colubrinus, P. flavoridis and L. muta venoms in a concentration range of 0.1–2.0 mg/mL. Then 20 μL of ALEXA488-Bgt in the binding buffer was added at a final concentration of 0.293 pM, and incubation was continued for 5 min more. The reaction was stopped by filtration on GF/C filters (Whatman, United Kingdom) presoaked in 0.5%-polyethyleneimine solution and washed with the cold binding buffer, 3 times by 200 μL. The amount of the bound ALEXA488-Bgt was determined with a FPM-01 fluorimeter (Kortek, Russia). The excitation light wavelength was 482 nm, and the signal was recorded at 538 nm. Nonspecific binding was determined in the presence of a 200-fold excess of α-cobratoxin. Each experiment was performed twice in four replicates.

The graphs were plotted using OriginPro 7.5 (OriginLab Corporation, United States).

Cell cultivation and transient transfection. Neuro2a mouse neuroblastoma cells and HEK293 human embryonic kidney cells were used for transfection. The cells were obtained from the Russian Collection of Cell Cultures (St. Petersburg, Russia). The cells were cultured in the DMEM medium (PanEco, Russia) containing 10% fetal serum (HyClone, United States), 5 mM L-glutamine, gentamicin (10 µg/mL), and amphotericin B (0.25 µg/mL). The cell culture was grown in an MCO-15AC incubator (5% СО2, 37°С; Sanyo, Japan).

Cell transfection was performed with Lipofectamine 2000 (Invitrogen, United States). The plasmids encoding muscle-type nAChR subunits were kindly provided by Professor V. Witzemann (Max Planck Institute for Medical Research, Germany). The transfection mixture contained 250 μL of serum-free DMEM, 2.5 μL of Lipofectamine 2000, 1 μg of plasmid DNA mixture from mouse muscle nAChR subunits (at a ratio of 2α1 : 1β1 : 1δ : 1ε), and 1 μg of DNA of calcium ion sensor Case12 (Evrogen, Russia). Calcium imaging experiments were performed 72 h after the transfection to achieve the optimal level of nAChR expression on cell surface.

Detection of changes in the intracellular calcium concentration in response to the G. saxatilis and M. colubrinus venoms. Genetically encoded Case12 sensor (Evrogen, Russia) was used to determine the intracellular content of calcium ions. The changes in intracellular calcium level were detected at 516 nm.

The images were recorded with a Planc N quartz lens (20×/0.40) using an Olympus XM-10 CCD camera (Hamamtsu, Japan). Real-time images were collected and stored using the Cell-A software (Olympus, Japan); ImageJ, OriginPro 7.5, and Microsoft Office Excel were used for further processing. The changes in cell fluorescence intensity were recorded with an XM10 video camera (Olympus, Japan) at an exposure time of 500 ms. The cell chamber volume was ~300 µL, which provided complete replacement of solution in the chamber within no more than 15 s.

Acetylcholine solution, 40 μM in 1 mL of an external solution (140 mM NaCl, 5 mM KCl, 2 mM CaCl2, 2 mM MgCl2, 10 mM HEPES, 10 mM glucose, pH 7.4), was added to the cells, which were then washed three times with the same solution. Then the cells were incubated with the venoms for 5 min, and 40 μM acetylcholine was added again. The results were recorded for each individual cell.

The changes in response of each individual cell were calculated by the formula:

$$100-({{{{A}_{{\text{e}}}}} \mathord{\left/ {\vphantom {{{{A}_{{\text{e}}}}} {{{A}_{{\text{k}}}}}}} \right. \kern-0em} {{{A}_{{\text{k}}}}}} \times 100),$$

where Ae is the amplitude of cell response to acetylcholine after 5-min incubation with the venom and Ak is the amplitude of cell response to acetylcholine in the preliminary experiment.

The results are presented as the mean ± standard error. The average percent change in the response was calculated on the basis of individual values obtained.

Competitive radioligand binding assay. For competitive binding experiments, the suspension of membranes of the T. californica electric organ (at a final concentration of 1.2 nM α-bungarotoxin-binding sites) or GH4C1 cells (0.4 nM α-bungarotoxin-binding sites) in 20 mM Tris-HCl buffer, pH 8.0, containing bovine serum albumin at a concentration of 1 mg/mL, was incubated for 90 min with different concentrations of the peptide. Radioactive [125I]-labeled α-Bgt (500 Ci/mmol) was added to a final concentration of 0.2–0.4 nM and incubated for 5 min more. The determination of nonspecific binding, sample filtration and bound radioactivity counting were performed as described [23].

Electrophysiological measurements. Electrophysiological measurements on Xenopus oocytes were performed as described [23]. The oocytes were preincubated with different concentrations of macoluxin for 5 min before its coapplication with acetylcholine. The reversibility of nAChR inhibition by macoluxin was tested by incubation of the receptor with macoluxin for 5 min. Then the oocytes were washed with the buffer and the current induced by 30 μM acetylcholine was measured after 5 min.

CONCLUSIONS

The analysis of venoms of several snake species from the families Viperidae and Lamprophiidae has shown for the first time that the venom of the rear-fanged cat-eyed snake M. colubrinus from the family Lamprophiidae possesses the capaciry to inhibit nAChR. The peptide macoluxin inhibiting the muscle-type nAChR was isolated from the venom of M. colubrinus by high-performance liquid chromatography, and its amino acid sequence was determined. The search of homologous amino acid sequences in the databases of known proteins has shown that the macoluxin sequence has a high degree of similarity (91% identity) with a fragment of the sequence of snake venom metalloproteinases. This metalloproteinase fragment has an α-helical structure and is localized on the surface of the protein globule. Macoluxin was obtained in preparative quantities by solid-phase peptide synthesis. The study of its biological activity has shown that it competes with α‑Bgt for the binding to the nAChR of Torpedo, with the IC50 value of 47 μM. The peptide at concentrations below 50 μM did not compete with α-Bgt for the binding with α7 nAChR. At concentrations within a range of hundreds of micromoles, macoluxin inhibited acetylcholine-induced currents in muscle-type nAChR. The effect of inhibition was easily reversible by 5-min washing.

Thus, we have identified and characterized for the first time the peptide inhibiting muscle-type nAChR in the venom of a rear-fanged snake. Our data indicate that venom proteins can act as precursors of peptide toxins with a different type of biological activity.