The World Health Organization recognizes twenty neglected tropical diseases that affect the whole world, but their devastating effects are felt mostly in tropical areas and impoverished communities, where access to hospitals and medicines is limited ^[1][2][1,2]. These neglected tropical diseases can be divided into infectious diseases (19 diseases) and non-infectious diseases, of which there is only one—snakebite envenoming ^[3][4][3,4]. The mortality rate of neglected tropical diseases is ca. 200 thousand people per year, and they cause more than 19 million disabilities annually. In the last decade, the World Health Organization and governmental agencies have carried out several campaigns and investments, significantly reducing the number of infectious neglected tropical diseases. Nonetheless, the impact of snakebite envenoming remains alarming, with 81–138 thousand deaths yearly, the total of which may equal the number of deaths caused by the other 19 neglected tropical diseases altogether [5]. The intricate chemical and bioactive composition of snake venom means that it is challenging [6] to find a cure for snakebite envenoming. The current state-of-the-art treatment involves administering antibodies purified from the plasma of hyper-immunized equines ^[7][8][7,8]. However, it is very costly, and it requires transportation and storage in cold containers, as well as inpatient administration due to frequent anaphylactic reactions [9]. These constraints mean that this therapy is unavailable in the remote regions of resource-poor countries and in poor communities where most snakebites occur ^[1][7][1,7].

Snake venoms are complex cocktails comprising tens to hundreds of components, of which >90% (w/w) are proteins and peptides ^[10][11][12][10,11,12]. The identity and abundance of components are diverse, and they depend on the snake species, gender, age, habitat, and prey, among other factors. Intraspecific variability is also significant. Venomous snakes belong to three families of front-fanged snakes, the Viperidae (vipers), Elapidae, and Atractaspidae, and a family of rear-fanged venomous snakes, the Colubridae; however, regarding the latter, not all members are venomous [13]. This work focuses primarily on viper venom, as vipers are estimated to be responsible for a substantial proportion of snakebite fatalities [14].

Despite the fact that viper venom is composed of highly diverse elements, three enzymatic toxin families stand out due to their toxicity and abundance: metalloproteinases (vvMPs), secreted phospholipase A2 enzymes (vvPLA2s), and serine proteases (vvSPs) [12]. These families frequently account for ca. 70% of the total venom weight. Other components that are often present, but in smaller amounts, are the C-type lectins, cysteine-rich secretory proteins, and L-amino acid oxidases [15] (for a more complete list, see Ref. [15]). vvMPs and vvPLA2s are present in more than 80% of the ca. 165 known viper venom proteomes [12].

vvPLA2 is the second most abundant protein family in viper venom on average, surpassed only by vvMPs. It is thus evident that vvPLA2s and vvMPs are prime drug target candidates for treating snakebite envenoming. vvPLA2s play many pathophysiologic roles; regarding the ones found in viper venoms, the most common is myotoxicity ^{[16][17][18][19][20]}[16,17,18,19,20]. Nevertheless, other usual effects caused by vvPLA2s include platelet aggregation, anticoagulation, hemolysis, cytotoxicity, and neurotoxicity ^[21][22][21,22]. They also participate in inflammatory processes and display bactericidal activities [23]. As a whole, viper envenomation produces highly harmful consequences, such as muscle breakdown, neuromuscular paralysis, tissue necrosis and amputations, bleeding disorders, thrombosis, kidney and heart failure, hypovolemic shock, and often death ^{[24][25][26][27]}[24,25,26,27], effects for which vvPLA2 is at least partially responsible.

vvPLA2s are small, secreted proteins of 13–15 kDa. Together with the human synovial fluid PLA2, they form the PLA2 group IIA ^[28][29][30][28,29,30]. Within this group, almost all enzymes have seven disulfide bonds, significant sequence identity, and a common fold (Figure 1). Rare exceptions within the viper family include the Gaboon viper (Bitis gabonica), the Rhinoceros viper (Bitis nasicornis) and Palm pit vipers  (Trimeresurus genus) vvPLA2s.

The vvPLA2 enzymes are further characterized by their isoelectric point. Accordingly, acidic vvPLA2s typically have an isoelectric point of 4.0–5.5 with many acidic, negative residues, and basic vvPLA2s have an isoelectric point of ca. 8.0 with primarily basic and positive protein surface residues. The isoelectric point influences the vvPLA2 toxicity, and as a consequence, basic enzymes are far more toxic than acidic ones. The underlying reason still needs to be clarified but is supposed to lie in the different affinities that the acidic and basic enzymes have for the cell membrane phospholipids and/or for the cell membrane protein receptors.

vvPLA2 enzymes cleave phospholipids at position sn-2, releasing fatty acids and lysophosphatidic acids. The corresponding mechanism of this hydrolysis reaction is only partially understood at the atomic level, occurring always in the presence of a Ca²⁺ cofactor. The latter is coordinated to Asp49 (bidentate), three carbonyl groups belonging to both a glycine and a tyrosine or phenylalanine residues located in the Ca²⁺-binding loop, and the substrate’s phosphate and sn-2 carbonyl groups (or two water molecules when in the unbound state). According to the postulated mechanisms, His48 deprotonates one water molecule, either directly (“single-water mechanism” hypothesis) or through a second bridging water molecule (“assisted water mechanism” hypothesis), generating a hydroxide ion that attacks the sn-2 carbon of the phospholipid (Figure 2A). The hydrogen bonding of His48 to the carboxylate of Asp99 stabilizes the positive form of His48, facilitating water deprotonation. Whether or not the nucleophilic water molecule is Ca²⁺-bound is unclear. Anyhow, the Ca²⁺ stabilizes the reaction’s transition state by coordinating the oxyanion generated at the sn-2 carbonyl site (and eventually the hydroxide nucleophile), subsequently decaying into a tetrahedral intermediate (Figure 2B). Following the ester bond cleavage, the hydrolyzed products—fatty acid and lysophospholipid—are released (Figure 2C). No evidence exists for protein conformational changes during the cycle [33].

Monomeric and dimeric vvPLA2s are the most common quaternary structures found via X-ray crystallography ^{[19][34][35][36]}[19,34,35,36], and they are supposed to be the most common in solution too.

Two different dimeric quaternary structures have been found in cristallo, named “conventional” and “alternative” ^[37][37] (Figure S1). Nevertheless, thwe researchers will refer to them here as “extended” and “compact” dimers, respectively, as this nomenclature better reflects the nature of their quaternary structure, and the fact that the buried area of these two structures is very different. The extended dimer of a PLA2-like protein (i.e., a non-enzymatic homolog of the vvPLA2 enzymes, as discussed in Section 3.6Section 1.3.6), from, e.g., B. pauloensis, buries an area of 1022 Ų (Ref. [38]), and it has few intermonomer contacts, which may turn the dimerization in water less robust. Nevertheless, it has phospholipid binding sites open to the solvent. The compact conformation of the same enzyme achieves a larger buried area of 1491 Ų (Ref. [39]) and establishes many more intermonomer contacts than the extended conformation. It seems more stable in water, even though each monomer occludes the binding site of the other (more examples in Table S1). Thus, a conformational change or dissociation may be needed to expose the binding site of a compact dimer to the cell membrane (Figure S2).

vvPLA2s and similar human pancreatic PLA2s are believed to bind the cell membrane as monomers in most cases [40]. Therefore, if a PLA2 forms dimers in solution, dimer dissociation should precede or happen concomitantly with membrane binding. vvPLA2s have more affinity for membrane regions richer in negative than zwitterionic phospholipids, particularly the most toxic basic isoforms. Insights from an anion-assisted dimer of pancreatic porcine PLA2 ^[41][42][41,42] estimate that PLA2 buries 30–40 phospholipids, which form salt bridges with the PLA2 basic residues [43]. Acidic vvPLA2s are generally catalytically more active in vitro but less toxic in vivo than the basic isoforms [44]. In contrast, most basic vvPLA2s induce several toxic effects ^{[17][26][37][45][46]}[17,26,37,45,46].

vvPLA2s contain an α-helical N-terminal region believed to be implicated in critical functions, such as membrane binding, and vital structural areas, such as the substrate-binding hydrophobic channel. X-ray structures of PLA2s from the Indian cobra (Naja naja) [47] and the European bee (Apis mellifera) [48] venoms and from human [49] and pig (sus scrofa) [50] secretory PLA2s, solved in the presence of transition-state analogs, showed the N-terminal region interacting with the phospholipid substrate (Figure S3). The invariant residues Leu2, Phe5, and Ile9 delimitate the substrate-binding cavity on one side. The contribution of the N-terminal region to the catalytic activity has been extensively studied in mammalian PLA2s, where the deletion of 8–10 residues of the N-terminus resulted in a nearly complete loss of enzyme activity ^[51][52][51,52]. Moreover, deleting the ten N-terminal residues made the enzyme bind to membranes with lower affinity and random, non-specific orientations [53]. Such studies underline the importance of the N-terminal region in regulating the PLA2 function by determining the strength of membrane binding and the productive orientation of PLA2 at the membrane surface and in contributing to possible structural changes in the enzyme during interfacial activation ^[54][55][54,55]. However, few studies regarding this topic were performed with vvPLA2. Nevertheless, studies involving a vvPLA2 from Crotalus atrox confirmed that deleting the ten N-terminal residues results in a loss of enzymatic activity as well as the dimeric structure of the native enzyme [56].

There are two types of PLA2 proteins in viper venom: the previously discussed Ca²⁺-dependent catalytically active enzymes (vvPLA2) and the non-enzymatic PLA2 proteins (Figure 3). The latter are named PLA2-homologs, Lys-49 PLA2, or PLA2-like proteins (the last designation will be used here) and are exclusively found in viper venoms as they have diverged from the former during evolution.

The main differences found between vvPLA2 and PLA2-like proteins are as follows:

i.

The residue at position 49, which can be Asp in the enzymes and Lys or, more rarely, Ser, Asn, Gln, or Arg, in the PLA2-like proteins);

ii.

The active site Ca²⁺ cofactor, which is only present in vvPLA2;

iii.

The sequence and fold of the Ca²⁺-binding loop;

iv.

The sequence of the C-terminal region [43]. Viper venoms generally contain a mixture of vvPLA2s and PLA2-like isoforms, whose synergistic action determines their function and pathological effects [58].

PLA2-like proteins have been extensively studied for their ability to disrupt the cell membrane, causing myotoxicity ^{[16][58][59][60][61][62][63][64]}[16,58,59,60,61,62,63,64]. A large body of studies recently reviewed in Ref. [57] identifies the C-terminal region (residues 105–117) in the mature Bothrops asper myotoxin II, Uniprot ID: P24605, or equivalently residues 115–129 when using the standard numbering (i.e., a common residue numbering for mammalian pancreatic and venom PLA2 enzymes [32]), composed primarily of cationic, hydrophobic, and aromatic residues, as the primary determinant for myotoxicity. Moreover, the same region is involved in hyperalgesia and inflammation. In several instances, peptides with the 115–129 sequence retain, partly or wholly, the bioactivity of the complete enzyme [60]. However, given the differences between the bioactivity of the total enzyme and the 115–129 peptide, studies suggest that the C-terminal region is central but not solely responsible for PLA2 toxicity, with other residues, such as K20, K36, and K38, also being crucial for the effect ^{[16][45][64][65]}[16,45,64,65]. Like vvPLA2, PLA2-like proteins induce local membrane perturbations, allowing Ca²⁺, K⁺, and ATP to be internalized through membrane crossing, leading to extreme cytotoxic events ^[66][67][66,67].

Researchers are developing small-molecule drug treatments against snakebite envenoming to overcome the limitations of the antibody-based treatment ^[68][69][70][68,69,70]. The vvPLA2s are a primary drug target because they are expressed in more than 90% of viper species ^[10][11][12][10,11,12] and are associated with drastic pathologic processes ^[21][71][72][21,71,72]. Moreover, it is known that competitive enzymatic inhibitors such as varespladib ^{[69][73][74][75][76][77][78][79]}[69,73,74,75,76,77,78,79] also inhibit PLA2-like myotoxicity ^[69][74][80][69,74,80].