Three-Finger Toxin

Three-finger toxins (3FTx) constitute a major superfamily of small, non-enzymatic proteins (57 to 85 amino acids, ~6.5 to 9.5 kDa) that dominate the venoms of elapid snakes—including cobras (Naja), king cobras (Ophiophagus), kraits (Bungarus), mambas (Dendroaspis), coral snakes (Micrurus), Australian elapids, and hydrophiine sea snakes—as well as rear-fanged colubrid snakes (Boiga, Spilotes). Named for their shared tertiary fold, every 3FTx monomer consists of three elongated antiparallel β-stranded loops ("Finger I, Finger II, and Finger III") projecting downward like three fingers of a planar hand from a compact, globular hydrophobic core cross-braced by four invariant disulfide bonds (plus an optional fifth disulfide in specific subclasses).

Despite conserving this rigid three-finger scaffold across hundreds of isoforms, the 3FTx superfamily provides one of the premier textbook examples of functional radiation (neofunctionalization) on a conserved structural fold. Through focal positive selection (d_N/d_S > 2.5) concentrated on the solvent-exposed β-hairpin fingertips while the disulfide-bonded knuckle undergoes strong purifying selection, 3FTx paralogs have diverged to attack an extraordinary array of physiological targets: postsynaptic nicotinic acetylcholine receptors (nAChRs) (short- and long-chain α-neurotoxins such as α-bungarotoxin and erabutoxin), neuronal α₃β₂ nAChRs (homodimeric κ-bungarotoxin), muscarinic GPCRs (mamba muscarinic toxins MT1–MT7), L-type Ca²⁺ channels (calciseptine), acid-sensing ion channels (ASICs) (pain-blocking mambalgins), acetylcholinesterase (fasciculins), β-adrenergic receptors, and phospholipid bilayers (cytolytic cardiotoxins / cytotoxins).

Three β-Hairpin Loops and the Invariant Four-Disulfide Knuckle

In high-resolution X-ray and NMR structures of canonical 3FTx proteins—such as the sea snake short-chain neurotoxin erabutoxin b (Laticauda semifasciata, 62 residues, PDB: 3EBX), α-bungarotoxin (PDB: 2ABX), and κ-bungarotoxin (PDB: 1KBA)—the polypeptide folds into two antiparallel β-sheets: a shorter two-stranded β-sheet forming Finger I (Loop I, β1–β2) and a larger three-stranded antiparallel β-sheet comprising Finger II (Loop II, β3–β4) and Finger III (Loop III, β5).

All three fingers emerge from an upper hydrophobic core ("palm/knuckle") locked together by four strictly conserved core disulfide bridges (numbered according to the canonical eight-cysteine framework: Cys1–Cys3, Cys2–Cys4, Cys5–Cys6, and Cys7–Cys8), reinforced by an invariant structural salt bridge between Arg39/Arg37 and Asp/Glu plus a buried aromatic Tyr25 (which forms a tight hydrogen bond with Glu38 to stabilize the central β-sheet):

1. Short-Chain (Type I) 3FTx (60–62 residues, 4 Disulfides): Possess only the four invariant core disulfide bonds (Cys1–Cys3, Cys2–Cys4, Cys5–Cys6, Cys7–Cys8). This group includes short-chain α-neurotoxins (erabutoxins, cobrotoxin), cobra cardiotoxins/cytotoxins, mamba fasciculins, calciseptine, and mambalgins.
2. Long-Chain (Type II) 3FTx (66–75 residues, 5 Disulfides — Finger II Extra Cys–Cys): Contain the four core disulfides plus a fifth disulfide bond located at the distal tip of Finger II (e.g., Cys29–Cys33 in α-bungarotoxin and α-cobratoxin, Cys27–Cys31 in dimeric κ-bungarotoxin) as well as an extended C-terminal tail. This fifth disulfide pinches the tip of Finger II into a cyclical helix-like turn that enables binding to neuronal α₇ homopentameric nAChRs (which 4-disulfide short-chain neurotoxins cannot block effectively) in addition to skeletal muscle nAChRs.
3. Non-Conventional / Ancestral (Type III) 3FTx (62–82 residues, 5 Disulfides — Finger I Extra Cys–Cys): Contain the four core disulfides plus a fifth disulfide bond located within Finger I (Loop I) (PDB: 1TGX, Naja kaouthia weak toxin WTX, and colubrid prey-specific toxins such as denmotoxin and irditoxin). As detailed below, this 10-cysteine Finger-I disulfide pattern is the ancestral plesiotypic state inherited directly from vertebrate Ly6/uPAR housekeeping proteins.

From GPI-Anchored Cholinergic Brake (LYNX1) to Injectable Neurotoxin

How did snakes evolve a venom weapon pre-configured to dock with picomolar precision onto vertebrate nicotinic acetylcholine receptors? Comparative genomics, intron–exon architecture, and structural biology have proven that snake venom 3FTx genes evolved directly from endogenous vertebrate Ly6/uPAR (LU-domain) housekeeping genes—specifically a family of brain-expressed neuromodulators represented in humans and mice by LYNX1 (Ly6/Neurotoxin-1), LYNX2, SLURP1 (mutated in Mal de Meleda palmoplantar keratoderma), and the complement membrane-attack inhibitor CD59:

1. Endogenous LYNX1 Function in the Mammalian Brain: In vertebrate neurons (including human cerebral cortex, hippocampus, and Purkinje cells), LYNX1 is a 10-cysteine (5-disulfide) three-finger protein anchored to the outer leaflet of the postsynaptic plasma membrane by a C-terminal glycosylphosphatidylinositol (GPI) lipid anchor. NMR structure determination of human water-soluble ws-LYNX1 (Lyukmanova et al., 2011; PDB: 2L03) revealed an exact structural superposition (backbone RMSD ~1.4 Å) with snake three-finger neurotoxins and κ-bungarotoxin (PDB: 1KBA). In the healthy brain, GPI-anchored LYNX1 binds allosterically alongside α₄β₂ and α₇ nAChRs to act as a physiological molecular brake that accelerates receptor desensitization, limits excitotoxic Ca²⁺ influx, and closes the juvenile critical period for visual cortex synaptic plasticity.
2. Gene Duplication and Loss of the Exon-3 GPI-Anchor Signal: Both vertebrate LYNX1/Ly6 genes and snake venom 3FTx genes share an identical three-exon / two-intron genomic organization with conserved phase-1 and phase-2 splice junctions. Early in the Caenophidia (advanced snake) radiation (~60–80 million years ago), an ancestral Ly6/LYNX-like gene underwent duplication and recruitment into the oral venom gland (Duvernoy's gland). A frameshift/premature stop mutation deleted the C-terminal hydrophobic GPI-anchor attachment signal peptide in Exon 3, liberating the protein from the cell membrane as a soluble, secretable 10-cysteine non-conventional 3FTx (retaining the ancestral 5th disulfide in Finger I).
3. Secondary Loss of the Finger-I Disulfide and Gain of the Finger-II Disulfide: Subsequent gene duplication in elapid snakes deleted the ancestral Finger-I cysteines—generating compact 8-cysteine (4-disulfide) short-chain 3FTx—from which a subset acquired two new cysteines at the tip of Finger II to create 10-cysteine (5-disulfide) long-chain α- and κ-neurotoxins (α-Bgt, κ-Bgt).

Once freed from its membrane anchor and amplified into tandem gene arrays (often 15 to 50 paralogous 3FTx copies per elapid genome), the three-finger scaffold diversified into at least seven distinct pharmacological classes:

• Prey-Specific Colubrid Dimers (Irditoxin & Denmotoxin): Rear-fanged colubrid tree snakes such as the brown tree snake (Boiga irregularis) and mangrove catsnake (Boiga dendrophila) express irditoxin (PDB: 3D32, a covalent heterodimer of two non-conventional 3FTx subunits linked by an interchain disulfide bond) and monomeric denmotoxin (PDB: 2H7Z), which block avian and lizard skeletal muscle nAChRs with 100- to 1,000-fold higher potency than mammalian nAChRs.
• Acetylcholinesterase Inhibitors (Fasciculins, PDB: 1FSS): Green and black mambas (Dendroaspis) express fasciculin-1 and -2, which use Finger I and Finger II to cap the peripheral anionic site (PAS) at the gorge entrance of acetylcholinesterase (AChE) (K_i ≈ 10 pM), sterically blocking acetylcholine entry and causing generalized muscle fasciculations (tremors).
• Sub-Type Selective Muscarinic GPCR Toxins (MT1–MT7): Mamba venoms also contain muscarinic toxins (MT7, PDB: 2VLW in complex with human M₁ mAChR PDB: 6WJC), which abandoned ion channels entirely to become the most selective allosteric ligands known for G-protein-coupled M₁–M₄ muscarinic acetylcholine receptors.
• L-Type Calcium-Channel Blockers (Calciseptine & FS2): Isolated from black mamba (Dendroaspis polylepis) venom, calciseptine (60 residues, 4 disulfides) selectively blocks cardiovascular L-type voltage-gated Ca²⁺ channels (Ca_V1.2) via an interaction site overlapping the 1,4-dihydropyridine antihypertensive drug pocket (nifedipine), inducing smooth muscle relaxation and cardiac hypotension.
• Analgesic ASIC Blockers (Mambalgins, PDB: 2MFA & Cryo-EM 7CFT): Discovered in 2012 by Diochot, Lazdunski, and Lingueglia in black and green mamba venoms, mambalgin-1, -2, and -3 bind the thumb domain of central and peripheral acid-sensing ion channels (ASIC1a and ASIC1b), trapping the proton-sensing channel in the closed state and producing analgesia as potent as morphine in inflammatory and neuropathic pain models without respiratory depression or opioid tolerance.
• Membrane-Lytic Cobra Cardiotoxins / Cytotoxins (CTXs, PDB: 1CDT): Accounting for 40% to 60% of cobra (Naja) dry venom weight, cytotoxins (cardiotoxins) replace the hydrophilic receptor-binding fingertips of neurotoxins with a continuous apolar hydrophobic patch across the tips of Fingers I, II, and III, bordered by a cationic belt of conserved lysine residues. Upon electrostatic attraction to anionic phospholipids (phosphatidylserine on plasma membranes or cardiolipin on the inner mitochondrial membrane), the three hydrophobic fingertips insert into the lipid bilayer, inducing toroidal pore formation, cardiomyocyte depolarization, systolic cardiac arrest, and severe local dermonecrosis.

How did deadly snake three-finger neurotoxins evolve from a normal brain protein found in humans?

Vertebrates—including humans—express an endogenous brain protein called LYNX1 (PDB: 2L03), a 10-cysteine three-finger protein tethered to the neuronal membrane by a GPI lipid anchor where it gently modulates nicotinic acetylcholine receptors (nAChRs) to prevent over-excitation. In early advanced snakes, a duplicated LYNX1/Ly6-family gene became expressed in the oral venom gland and suffered a mutation in Exon 3 that deleted its C-terminal GPI membrane anchor, converting a tethered cholinergic regulator into a soluble, injectable neurotoxin already shaped to bind nAChRs.

Why can long-chain α-neurotoxins (like α-bungarotoxin) block neuronal α₇ receptors, whereas short-chain α-neurotoxins cannot?

Both short-chain (60–62 aa, 4 disulfides) and long-chain (66–75 aa, 5 disulfides) α-neurotoxins bind with picomolar affinity to skeletal muscle ((α₁)₂β₁δε) nAChRs. However, long-chain α-neurotoxins possess an extra fifth disulfide bond at the distal tip of Finger II (e.g., Cys29–Cys33 in α-bungarotoxin) and a longer C-terminal tail. This cyclical Finger-II tip forms specific hydrophobic and π–cation contacts inside the homopentameric α₇–α₇ interface that flat Finger-II tips of 4-disulfide short-chain neurotoxins cannot establish.

How can mambalgins and fasciculins have completely different effects in the body if they share the exact same three-finger fold?

The four core disulfide bridges in the upper "knuckle" of a three-finger toxin (3FTx) lock the three β-hairpin loops into a stable, protease-resistant scaffold regardless of which amino acids sit on the outer loop surfaces. By rapidly mutating the surface residues on Fingers I, II, and III (d_N/d_S > 2.5), mamba paralogs evolved one surface that plugs the catalytic gorge of acetylcholinesterase (fasciculin-2, causing muscle tremors) and another surface that clamps the thumb domain of ASIC1a proton channels (mambalgin-1, blocking pain signaling).

1. Fry, B. G., Wüster, W., Kini, R. M., Brusic, V., Khan, A., Venkataraman, D., & Rooney, A. P. (2003). "Molecular evolution and phylogeny of elapid snake venom three-finger toxins." Journal of Molecular Evolution, 57(1), 110–129. https://doi.org/10.1007/s00239-003-2461-2
2. Miwa, J. M., Ibañez-Tallon, I., Crabtree, G. W., Sánchez, R., Šali, A., Role, L. W., & Heintz, N. (1999). "lynx1, an endogenous toxin-like modulator of nicotinic acetylcholine receptors in the mammalian CNS." Neuron, 23(1), 105–114. https://doi.org/10.1016/S0896-6273(00)80757-680757-6)
3. Kini, R. M., & Doley, R. (2010). "Structure, function and evolution of three-finger toxins: mini proteins with multiple targets." Toxicon, 56(6), 855–867. https://doi.org/10.1016/j.toxicon.2010.07.010
4. Diochot, S., Baron, A., Salinas, M., Douguet, D., Scarzello, S., Dabert-Gay, A. S., ... & Lazdunski, M. (2012). "Black mamba venom peptides target acid-sensing ion channels to abolish pain." Nature, 490(7421), 552–555. https://doi.org/10.1038/nature11494