Catalytically Active Snake Venom PLA₂ Enzymes: An Overview of Its Elusive Mechanisms of Reaction

Abstract

Snake venom-secreted phospholipase A₂ (svPLA₂) enzymes, both catalytically active and inactive, are a central component in envenoming. These are responsible for disrupting the cell membrane’s integrity, inducing a wide range of pharmacological effects, such as the necrosis of the bitten limb, cardiorespiratory arrest, edema, and anticoagulation. Although extensively characterized, the reaction mechanisms of enzymatic svPLA₂ are still to be thoroughly understood. This review presents and analyses the most plausible reaction mechanisms for svPLA_2, such as the “single-water mechanism” or the “assisted-water mechanism” initially proposed for the homologous human PLA₂. All of the mechanistic possibilities are characterized by a highly conserved Asp/His/water triad and a Ca²⁺ cofactor. The extraordinary increase in activity induced by binding to a lipid–water interface, known as “interfacial activation,” critical for the PLA₂s activity, is also discussed. Finally, a potential catalytic mechanism for the postulated noncatalytic PLA₂-like proteins is anticipated.

1. Introduction to Snake Venom

1.1. Epidemiology and Snake Venom Composition: A Brief Account

Snakebite envenomation is a significant public health concern worldwide, particularly in resource-poor regions of tropical and subtropical countries. It is associated with a broad spectrum of pathophysiological effects resulting in high morbidity and mortality. Every year 81 000–138 000 people die from snakebites, and over 400 000 suffer permanent sequelae, such as amputations. The WHO thus recognizes snakebite as the deadliest neglected tropical disease.¹⁻⁵ Administration of animal-derived antivenom remains the more viable available therapy for treating snake envenomation,^6,7 and early intervention after envenoming is crucial for preventing venom-induced toxicity.⁶ It is particularly relevant for avoiding myotoxicity, a severe condition sometimes coupled with neurotoxicity,^8,9 consequences of envenomation caused by several snakes of the Elapidae, Viperidae, Atractaspididae, and sometimes Colubridae families.⁸⁻¹⁰ Nevertheless, the antivenom treatment has significant drawbacks: it is costly, needs inpatient administration by trained physicians, refrigerated transport and storage, and its efficacy is limited to the species used in the animal immunization and frequently causes anaphylactic reactions.⁵ Alternatives based on small molecule inhibitors of central venom toxins, such as the svPLA₂ discussed in this Perspective, are being developed to overcome some of these limitations.^1,11

Snake venoms are cocktails of bioactive molecules. The venom of each species is unique, and intraspecific variations are common.^1,12 The snake venoms are mixtures of tenths to hundreds of different components, from which proteins and peptides generally constitute more than 90% of the dry weight.¹ The enzymes secreted phospholipase A₂ (svPLA₂s), metalloproteases, and serine proteases are the most abundant and relevant components in viperid snakes, but there are immense variations.¹ In Elapid venoms, the most abundant and relevant components are three-finger toxins and svPLA₂, but the list of variants and exceptions is vast.^1,13,14

Therefore, snake venoms are a rich source of svPLA₂ enzymes (EC 3.1.1.4).^1,10 This enzyme family of enzymes is expressed in almost all venomous snakes,⁸ with Elapidae and Viperidae families exhibiting the highest concentrations and Colubridae (usually nonvenomous) displaying the lowest.¹⁵

1.2. PLA₂ Enzymes in Snake Venom

The PLA₂ superfamily is a broad class of esterases defined by their ability to catalyze the hydrolysis of membrane glycerophospholipids at the sn-2 position (Figure 1). The principal contribution of the reaction catalyzed by PLA₂ is the permeabilization and disruption of the cell membrane integrity, leading to an uncontrolled influx of extracellular molecules, such as Ca²⁺ ions, which trigger a set of events that lead to cell death.^1,16 In addition, the reaction catalyzed by PLA₂s releases free fatty acids, such as arachidonic acid and lysophospholipids, which are precursors for several signaling molecules involved in biological processes.^10,17,18 For example, arachidonic acid can be converted into prostaglandins and leukotrienes, contributing to the mediation of inflammatory processes and pain. On the other hand, the lysophospholipids can be acetylated to platelet-activating factors, which have diverse physiologic roles in the immune system.^17,19

Figure 1.

The chemical reaction catalyzed by PLA₂s. These enzymes act at the sn-2 position of glycerophospholipids and hydrolyze the ester bond releasing a lysophospholipid and a free fatty acid. R1 and R2 correspond to the fatty acid tails.

Over the last decades, significant progress has been made in clarifying the role of the broad PLA₂ superfamily in many biological functions, particularly in mammals, focusing on human enzymes.¹⁸⁻²¹ These include membrane remodeling, signal transduction, inflammation, antimicrobial defense, metabolism, platelet activation, and cell signaling.^17,18,20,22 The wide range of biological effects has attracted the interest of many scientists, as PLA₂s are obvious therapeutic targets for developing pharmaceuticals.^1,17 However, while the mammalian PLA₂s enzymes are often nontoxic,²¹ the svPLA₂s, which are the focus of this Perspective, induce a vast array of toxic effects, including cytotoxicity, hemotoxicity, proinflammatory, (anti)coagulant, and hypotensive effects, besides the above-mentioned myotoxic and neurotoxic effects. These toxic effects may result from catalytic and/or noncatalytic activities.^7,21,23−26 Although many PLA₂s discovered in humans have undergone several structural and functional studies, the same does not apply to svPLA₂, for which our knowledge of the reaction mechanisms is still limited. As a result, this work focuses on the svPLA₂ potential reaction mechanisms, discussing the current proposals for their chemistry and the arguments supporting each, and drawing comparisons with the proposed reaction for the human PLA₂.

Based on structural and functional features, the PLA₂ superfamily is divided into four principal categories: the Ca²⁺-dependent secreted, further subdivided into 17 groups,²⁷ the Ca²⁺-dependent cytosolic, the Ca²⁺-independent, and the lipoprotein-associated (LpPLA₂s) phospholipase A₂ enzymes.^17,28 All svPLA₂ are of the secreted category. The details concerning the remaining types and groups fall out of the scope of this Perspective and can be found in several excellent works on this matter.^18,22

The svPLA₂ are small extracellular proteins with a molecular weight of 14–18 kDa, 120–135 amino acid residues, and 6–8 disulfide bridges that contribute to their high degree of stability and a pH optimum at 7.^29,30 In addition, these enzymes are characterized by a His48/Asp99 dyad at the active site with a Ca²⁺-binding loop and the requirement of mM Ca²⁺ for catalytic activity.^22,27,28 Besides snake venom, secreted PLA₂ are found in scorpion and bee venom, among other animal venoms, and in several body fluids as nontoxic enzymes, including blood plasma, pancreatic secretions, seminal fluid, or tears.^19,27 svPLA₂ from elapid and viperid snakes share six disulfide bonds and an additional one in a different location on each.²⁹

A group numbering system was established based on differences in disulfide bonding patterns, amino acid sequences, molecular weight, and loop insertion²⁷ (see Dennis et al.²⁹ for a thorough overview of sPLA₂ classification and history). According to it, snake venoms are divided into two groups of svPLA₂s, group IA (svPLA₂-IA), found in snakes of the Elapidae family, and IIA (svPLA₂-IIA), found in snakes of the Viperidae family.^15,18,31 The third variant of svPLA₂ has been found in rear-fanged snakes and classified as group IIE.¹⁵ However, information about this group is still scarce. Thus, this review will not discuss the group IIE in detail. For example, the venom of the Indian spectacled cobra (Naja naja, Elapidae family) belongs to group IA, while the Indian Russell’s Viper svPLA₂ (Daboia russelii, Viperidae family) belongs to group IIA.³² Traditionally, the svPLA₂ of the Gaboon viper (Bitis gabonica) is placed separately as the only member of the group-IIB svPLA₂ due to having only six disulfide bonds. However, other vipers, such as the rhinoceros viper (Bitis nasicornis) and the Saharan horned viper (Cerastes cerastes), share this characteristic and should be placed in the same svPLA₂-IIB group.¹⁵ The nonvenomous mammalian secreted sPLA₂ that are most similar to the svPLA₂ are the pancreatic sPLA₂ (group IB), which have a similar disulfide bond pattern to that of the svPLA_2-IA and the synovial-specific sPLA₂ isolated from arthritic synovial fluids and platelets, classified as GPLA₂-IIA, given their disulfide similarity with the svPLA₂s -GIIA.^15,27,33

The svPLA₂-IAs comprise a single polypeptide chain with 115–125 residues and six disulfide bridges, with an additional one between Cys11 and Cys71 (Renetseder et al.³⁴ amino acid numbering). This type of enzyme is ubiquitous in the venom of elapid snakes and accommodates the so-called “elapid loop,” an insertion of 2–3 residues that connects the second α-helix and the β-wing.^15,22 As an example, the mammalian pancreatic PLA₂ (PLA₂-IB) also possesses a loop with five additional amino acid residues at positions 62–67, named as the pancreatic loop.^35,36

The svPLA₂-IIA group includes both snake venom and mammalian nonpancreatic PLA₂. These enzymes comprise 120–125 amino acid residues and six disulfide bonds, with an additional one between Cys133–Cys50. Interestingly, svPLA₂-IIA enzymes lack the elapid loop present in svPLA₂-IA. They display, however, an extension of 5–7 amino acid residues at the C-terminal region.^35,36

2. The Structure of svPLA₂s

Over 200 nonredundant, complete, and reviewed sequences of svPLA₂s can be found in the UniProtKB Database (accessed in March).³⁷ Their primary structure shows a significant percentage of sequence identity among species, and the X-ray structures deposited in the Protein Data Bank (PDB) (e.g., snake venoms, humans, and bovines) revealed that svPLA₂ groups have a markedly similar architecture.^27,29,34,38 A sequence alignment denoting the conservation of the catalytic center of different species of secreted PLA₂ is shown in Figure 2, and a sequence identity matrix is presented in Supporting Information (SI), Table S1.

Figure 2.

Sequence alignment of six secreted phospholipase A₂ from different sources: human synovial (UniProtKB AC: P14555), bovine pancreatic (UniProtKB AC: P00593), B. asper (UniProtKB AC: P20474), E. carinatus (UniProtKB AC: Q7T3S7), N. atra (UniProtKB AC: P00598) and N. sputatrix (UniProtKB AC: Q92085). Overall, highly conserved residues can be found in the active site, Ca²⁺ binding loop and disulfide bonds. Amino acids residues that have an identity threshold above 40% are colored according to the ClustalX color scheme: hydrophobic (blue), positive charge (red), negative charge (magenta), polar (green), cysteines (pink), glycines (orange), prolines (yellow), aromatic (cyan), and unconserved (white). The numbering shown is that from Renetseder et al.³⁴ Green triangles indicate the locations of the residues involved in the calcium-binding, while yellow stars indicate those involved in the active site.

The svPLA₂s conserved tertiary structure is mainly characterized by an N-terminal α-helix (α1), two disulfide-connected α-helices (α2 and α3) where the catalytic dyad is placed, a double-stranded antiparallel β-sheet (β-wing), a Ca²⁺-binding loop, and a flexible C-terminal loop (Figure 3). Their folding is also highly similar to the bovine pancreatic and human synovial PLA₂ (SI, Figure S1).

Figure 3.

| (A) Structure of the Chinese cobra svPLA₂-IA (PDB 1POA) and (B) the Indian saw-scaled viper svPLA₂-IIA (PDB 1OZ6). The active site residues (His48, Asp49, Tyr52, Tyr73, and Asp99) are shown as green sticks, the Ca²⁺ as an orange sphere, and the disulfide bonds as yellow lines. N- and C-terminal regions are also identified. The similarity in the folding is evident. The PyMOL³⁹ molecular graphics software package was used to generate the representations.

In both svPLA₂ groups, the space between the two antiparallel α-helices is composed of highly hydrophilic solvent-exposed amino acid side chains, and hydrophobic residues are orientated toward the protein’s core. However, the polar residues comprising the catalytic dyad and its hydrogen bond network (His48, Asp49, Tyr52, Tyr73, and Asp99) are buried in the protein core (Figure 4, left).^{22,35,40−42} svPLA₂s contain an N-terminal α-helix (α1) followed by a short helix. The hydrogen-bonding network formed by these N-terminal α-helices builds a “hydrophobic channel” made by the highly hydrophobic side chains of Leu2, Phe5, Ile9, and Trp19 residues (more common in svPLA₂-GI) (Figure 4, right) that enhance phospholipid binding while also acting as a shield for His48/Asp99 residues against the surrounding solvent.^31,41,42

Figure 4.

| (left) Ball and stick representation of the residues involved in the catalytic network and respective hydrogen-bonding (dashed lines). (right) Surface representation of the GIA-PLA₂ isolated from N. atra (PDB 1POA) and stick representation of the residues that constitute the hydrophobic channel. The PyMOL molecular graphics software package was used to generate the representations.

Moreover, it consists of a Ca²⁺-binding loop (Figure 3)²² characterized by the carbonyl groups of Tyr28, Gly30, and Gly32, which, together with the β-carboxyl group of Asp49 coordinate the cofactor (Figure 3).¹⁵ The Tyr28 residue is crucial in binding Ca²⁺ due to the electrostatic interactions between its hydroxyl group oxygen and the Gly35, which result in a greater capacity to bind Ca²⁺ and, thus, a more potent catalytic activity.^22,41 The distance between the hydroxyl oxygen of Tyr28 and the amino group of Gly35 (around 3.5 Å) is conserved in almost all Asp49 PLA₂s. This interaction provides the Ca^2+-binding loop with higher structural stability and a better conformation for metal binding. In agreement, the Asp49Ser svPLA₂ from E. carinatus shows a distorted Ca²⁺-binding loop due to the absence of this interaction.⁴³

This loop is followed by the second α-helix (α2), which binds to antiparallel β-sheets cross-linked by disulfide bonds (β-wing region). Following this region is the α-helix 3 (α3), which binds to an exceptionally flexible region, the C-terminal loop, that is believed to be involved in the biological effects of these toxins^22,27 and allows it to change its conformation and interact with natural lipids.^22,40 The crystal structures of svPLA₂-GIA from the Chinese cobra (Naja atra) and svPLA₂-GIIA from the Indian saw-scaled viper (Echis carinatus) venom were used to illustrate these structural characteristics (Figure 3). In addition, structures of bovine pancreatic and human synovial PLA₂ are also shown (SI, Figure S1) to emphasize the above-mentioned similar architecture.

3. The Reaction Mechanisms of svPLA₂

3.1. Introduction

Two catalytic mechanisms have been proposed for the family of the secreted PLA₂s: the “single-water mechanism”^30,44 and the “assisted-water mechanism”.⁴⁵ The proposals resulted from analyzing X-ray structures of several PLA₂ from different organisms bound to inhibitors, substrate- and transition-state analogues,⁴⁶⁻⁵⁰ results from mutagenesis, and known chemical mechanisms of related enzymes. The crystal structures of about 40 group-I and group-II PLA₂ from several sources have been determined and deposited in the PDB,⁴¹ a curated selection of which is given in SI, Table S2. Some have been crystallized with ligands bound in the active site (holo form), whereas others were in their apo form (SI, Table S2).

The X-ray structures of several svPLA₂s–inhibitor complexes, three with a transition-state analogue (PDBs 1POB,⁴⁸1POE,⁴⁹ and 1POC(50)) and one with a substrate-derived amide analogue (PDB 5P2P(46)) were fundamental to investigate the mechanisms of reaction.³⁵ All of these structures share key properties involved in the catalytic process. According to crystallographic, biochemical, and site-directed mutagenesis studies,⁵¹ the active site (CCXXH₄₈D₄₉XC) and the Ca²⁺-binding loop (GCY₂₈CG₃₀XG₃₂GXG) motifs are the most conserved and relevant regions of the PLA₂ protein.¹⁵

Mutagenesis also provided fundamental mechanistic information. The most significant is the single-point mutations D49K and D49E, G30S, and H48Q.⁵² Results showed a loss of Ca²⁺ binding and a subsequent loss of enzymatic activity caused by the D49K substitution. Contrarily, the D49E mutant retained the ability to bind Ca²⁺, but this was insufficient for the reaction to occur. This phenomena could be because, with this substitution, the distance between the substrate’s ester bond and the Ca²⁺ ion increases, making it unable to stabilize the tetrahedral intermediate efficiently. These findings unequivocally demonstrate that the side chain carboxylate of Asp49 plays a crucial role in the enzyme’s capacity to bind Ca²⁺.^53,54 Moreover, the H48Q mutation reduced the enzymatic activity dramatically, and the G30S mutation affected the binding of substrate and Ca²⁺. Thus, the mutagenesis tests pointed out Asp49 and G30 as necessary for Ca²⁺-dependent phospholipid binding and His48 for their hydrolysis, crucial for catalysis.⁵²⁻⁵⁵

Finally, it was concluded that the conserved His48/Asp99 dyad plus a water molecule constitute a “catalytic triad” similar to the one observed in serine proteases and serine esterases, with the water molecule playing the role of the serine.⁵⁶ In analogy with the serine proteases/esterases, the His48 Nε2 atom appears hydrogen-bonded to the carboxylate Oδ1 atom of Asp99 in X-ray structures.^30,33,41,57 The hydrogen bond raises the pK_a of the His48, increasing the basicity of its Nδ1 atom.³³ Moreover, the pK_a of the His48 Nδ1 drops from ∼6.5 to 5.5 in the presence of Ca²⁺ ions.⁴²

Verheij et al. postulated a catalytic mechanism for the secreted PLA₂ known as the “single-water mechanism″⁴⁴ based on these findings.^30,42 Additional validation of the proposed mechanism came later from Scott et al.³⁰ based on the structures of secreted PLA2:transition-state analogue complexes.

Later, the study of a bovine pancreatic PLA₂-GIB cocrystallized with the transition-state analogue 1-hexadecyl-3-(trifluoroethyl)-sn-glycero-2-phosphomethanol (MJ33, PDB 1FDK(47)) led to the discovery of a second, Ca²⁺-bound, water molecule, which was proposed to play the role of the nucleophile.⁵⁸ Subsequent studies led Yu et al. to propose an alternative catalytic mechanism, the “assisted-water mechanism”.⁴⁵ In this context, the initial enzyme–substrate complex involves two water molecules at the active site, one that is assisted and Ca²⁺-bound (Ca²⁺-inner coordination sphere). It performs a nucleophilic attack on the substrate and an additional water (Ca²⁺-outer coordination sphere) that acts as a bridge between the attacking water and the His48 base, helping to overcome the considerable distance between the Ca²⁺-bound water proton and the basic His48 Nδ1.

Besides these two mechanisms, it is possible to envisage further alternatives based on the catalytic function of each residue in the mechanisms mentioned above and on known mechanisms from similar enzymes, which involve, for example, a bulk hydroxide ion instead of water as a nucleophile, or the inclusion of additional water in the active center. These mechanisms will also be discussed below.

3.2. The Single-Water Mechanism

According to Verheij (Figure 5), this mechanistic proposal starts with the Ca²⁺ ion hepta-coordinated by a bipyramidal pentagonal cage of oxygen atoms (Figure 5B),^30,35,44 constituted by the two carboxylate oxygens of the Asp49 side chain, three backbone carbonyl oxygen atoms of the Ca²⁺-binding loop (Tyr28, Gly30, and Gly32),^41,42 and two water molecules.³⁵ The water molecules are displaced upon substrate binding, specifically by the phospholipid phosphate and sn-2 carbonyl oxygens.²⁹ The former coordination strengthens the substrate’s binding, whereas the latter lowers the reaction activation barrier.³⁰

Figure 5.

General scheme of the single-water mechanism. (A) The Ca²⁺ coordination shell and the catalytically relevant residues. Their representation will be simplified in the following schemes for simplicity: (B) Verheij proposal,⁴⁴ (C) our proposal, with one calcium-bound water. Step 1: His48 abstracts a proton from the incoming water, which initiates a nucleophilic attack on the sn-2 carbonyl carbon of the substrate. Step 2: The produced tetrahedral intermediate oxyanion collapses, eliminating the alkoxyl group, which deprotonates His48. Step 3: Products release; His48 is stabilized by Asp99, which additionally forms hydrogen bonds with Tyr52 and Tyr73. The oxyanion hole that stabilizes the transition state after the nucleophilic attack is formed by the backbone HN group of Gly30 and the Ca²⁺ ion.

Subsequently, a structurally conserved active-site water molecule is polarized and deprotonated by the His48 Nδ1 atom (Figure 5A), itself polarized by hydrogen bonding of its Nε2 atom to the Asp99 Oδ1 atom. The latter residue stabilizes the cationic form of His48, permitting the deprotonation of water by a formally much less basic species. The formation of a hydroxide ion triggers a nucleophilic attack on the sn-2 carbon. This step leads to the formation of a tetrahedral oxyanion intermediate, which is thought to be rate-limiting.^29,30,59 Moreover, the tetrahedral intermediate’s oxyanion is stabilized by hydrogen bonding to the backbone amine of Gly30 and coordination with the Lewis acid Ca²⁺. The Ca²⁺ cofactor is thought to play a similar role to the Zn²⁺ ion in carboxypeptidases.^42,60 After the first step, the tetrahedral intermediate collapses by transferring the His48 δNH⁺ proton to the oxyanion of the leaving group.^30,42 Once the products are released, three water molecules migrate into the active site, from which two coordinate the Ca²⁺ ion and the third replenishes the active cycle for nucleophilic attack of the subsequent turnover.^30,41

Based on the mechanism of action indicated above, we infer a modified version of the postulated Verheij mechanism (Figure 5B). In our perspective, this new proposal makes more chemical sense and fits the available experimental data as it includes the fundamental catalytic role of the Ca²⁺ cofactor and the residues His48 and Asp99, which are known to be crucial for the mechanism. However, we agree that our modified version of the postulated Verheij mechanism is not directly supported by the X-ray structures available. The issue with the X-ray structures is that few have the Ca²⁺ cofactor cocrystallized and few have a substrate/transition state analogue cocrystallized. Taking the two premises together, around five structures are available (SI, Table S2, holo structures). In those structures, the Ca²⁺ is heptacoordinated to residues 28, 30, 32, and 49 (double coordination) and the ligand phosphate (double coordination). As the coordination number of the Ca²⁺ is seven, there is no place for a water molecule. Nevertheless, in bulk solution, the very abundant water molecules can easily replace one of the carbonyl groups of residues 28, 30, or 32, as water is a better ligand for Ca²⁺. Thus, in this alternative, the initial enzyme–substrate complex involves a Ca²⁺-coordinated water molecule instead of a doubly coordinated Asp49. Such a configuration will be more suitable for the reaction to progress as binding to the Ca²⁺ ion lowers the pK_a of the bound water molecule (in the same way Zn²⁺ does in snake venom metalloproteinases^1,61 facilitating its deprotonation). Furthermore, the binding of the water molecule with the consequent displacement of a coordination position of Asp49 oxygen can occur through the “carboxylate-shift” mechanism, a well-known and studied phenomenon in Zn²⁺ coordination shells.⁶² A similar mechanism has also been observed for sulfur bound to molybdenum cofactors and denominated as “sulfur shift”.⁶³ Alternatively, one of the critical Ca²⁺-binding loop residues may undergo a configurational change, uncoordinating the Ca²⁺ ion. The reaction’s subsequent steps are carried out the same way as in the Verheij proposal.⁴⁴

3.3. The Assisted-Water Mechanism

The assisted-water mechanism was proposed by Yu et al.⁴⁵ and is represented in Figure 6. In this proposal, the Ca²⁺ ion coordinates a water molecule and lowers its pK_a. This process is assisted by the hydrogen bond established with the more basic electron pair of Asp49. The nucleophilic Ca²⁺-bound water is hydrogen-bonded to a second water molecule, which, in turn, is hydrogen-bonded to His48. This hydrogen-bonding network overcomes the considerable distance between Ca²⁺ and the His48 Nδ1 atom, corresponding to ca. 6.2 Å.

Figure 6.

Schematic representation of the assisted-water mechanism proposed by Yu et al.⁴⁵ Step 1: His48, which acts as a general base catalyst, abstracts a proton from the second water, which deprotonates the calcium-bound water molecule. This leads to the nucleophilic attack by the calcium-bound water on the carbonyl carbon of the substrate and the formation of the tetrahedral intermediate. Step 2: The departing alcoholate leaving group is protonated by the second water, which is itself protonated by His 48. Step 3: Collapses and the products are released.

According to Yu and co-workers, in the first step of the mechanism, the catalytic water molecule deprotonates the second bridging one, itself deprotonated by the Nδ1 atom of the His48, thereby facilitating the reaction (Figure 6). Subsequently, the generated Ca²⁺-bound hydroxide ion water makes a nucleophilic attack on the substrate sn-2 carbon atom. This leads to the formation of a tetrahedral intermediate with a Ca²⁺-coordinated oxyanion. During the fallout of the tetrahedral intermediate, His48 protonates the bridging water molecule, which, in turn, protonates the departing alkoxy oxygen.⁴⁵

3.4. The Direct Hydroxide Attack Mechanism

Finally, the nucleophilic attack may occur through a hydroxide ion from the bulk water coordinated by the metal cofactor. Divalent metal ions have a high affinity for hydroxide ions. At physiologic pH and temperature (7.4 and 37 °C, respectively), where the hydroxide concentration is ∼56 × 10⁷ lower than water, the entropic cost has been estimated to be around 12.2 kcal·mol^–1. However, there are known cases where the binding free energy of the hydroxide ion to enzymes divalent cation compensates almost entirely for the high entropic cost of exchanging water by hydroxide into the Ca²⁺ coordination shell, and the lower barrier they provide compensates for the cost of the lower abundance of hydroxide-bound enzymes (2.7 kcal·mol^–1).^64,65

In this scenario, the Ca²⁺-bound hydroxide attacks the sn-2 carbon of the substrate, forming the tetrahedral intermediate (Figure 7). In addition, the Ca²⁺ cofactor stabilizes the C–O^– bond by coordinating it. The intermediate then collapses, reconstructing the C=O bonded to the metal ion, resulting in the cleavage of the bond and departure of the alkoxide group (RO^–). The latter may bind strongly to the Ca²⁺ cofactor, replacing the ionic interaction lost with the hydroxide neutralization. However, the high basicity of the sn-2 oxygen makes it a poor leaving group

Figure 7.

Schematic representation of the hydroxide direct nucleophilic attack mechanism; Step 1: The hydroxide nucleophile attacks at the electrophilic C of the ester C=O leading to the formation of the tetrahedral intermediate. Step 2: The intermediate collapses, reforming the C=O. Step 3: Departure of the alkoxide leaving group, RO^–.

Nevertheless, the oxyanion can be easily protonated by the much more acidic fatty acid carboxylate, generating two suitable leaving groups. Alternatively, an incoming water molecule can bind the active site and be deprotonated by the sn-2 oxygen, easing the elimination of the lysophospholipid and generating the hydroxide ion for the next catalytic cycle. In this variant, a bulk hydroxide ion binding would be needed only for the “initiation” turnover, with the nucleophile for the subsequent turnovers generated by the reaction product.

4. The Catalytic Mechanism of the Noncatalytic PLA₂ Proteins

In addition to the classical sPLA₂ enzymes, viperid venoms possess several toxins designated as “svPLA₂-like” proteins or “Lys49 svPLA₂ homologues” (SI, Figure S2). These proteins share a pervasive sequence identity and folding similarity to the svPLA₂ enzymes. However, they are essentially catalytically inactive.⁶⁶ The fact that svPLA₂-like proteins exist exclusively in viperid venoms indicates that they emerged after the Elapidae and Viperidae family divergence.¹⁵ Intriguingly, while having extremely limited or no enzymatic activity, Lys49-PLA₂-like enzymes can elicit several pharmacological effects, such as myotoxicity, cytotoxicity, hyperalgesia, and edema-inducing activities, irrespective of their catalytic activity.^67,68

PLA₂-like proteins have the Asp49 residue substituted by a Lys or, more rarely, by a Ser, Asn, or Gln residue. Unfortunately, there is not enough data about the Ser/Asn/Gln variants to draw a picture of their catalytic possibilities. It is tempting to speculate that at least the Ser variant might retain some activity, as the Ser-His-Asp constitutes the classical catalytic triad of serine esterases.⁵⁶ However, the correct substrate position and a proper oxyanion hole are mandatory for the reaction, but there is not enough data or model about this system.

As mentioned before, Asp49 is crucial for Ca²⁺ binding and, thus, for the catalytic mechanism (Figures 4 and 5). In addition, the Ca²⁺-binding loop has an open conformation due to the Tyr28Asn mutation, which induces a more extended Asn28–Gly35 interaction.⁶⁹ This loop conformation further impairs the coordination of the Ca²⁺ cofactor.⁶⁹ Thus, the loss of the productive Ca²⁺-binding loop conformation might be a reason why the Lys49 sPLA₂ cannot bind the Ca²⁺ ion.^69,70 The lack of Ca²⁺, essential for stabilizing the tetrahedral intermediate, was believed to be why these proteins cannot perform the catalytic cleavage of phospholipids.^69,71 This theory was backed by structural studies on the svPLA₂-like enzyme from A. piscivorus piscivorus, which revealed that the Lys49 Nε atom was placed in the Ca²⁺ atom position from Asp49-svPLA₂.⁷¹⁻⁷³ However, in vitro assays have demonstrated that these variants exhibit catalytic activity, albeit they are minimal.^74,75 Nevertheless, this observation was contested by others, which attributed the residual catalytic activity of PLA₂-like proteins to insufficient protein purification and contamination with PLA₂ enzymes.^53,73,76

However, experiments by Pedersen et al.⁷⁴ indicated that PLA₂-like proteins might cleave phospholipids but cannot release the fatty acid product and suggested a covalent binding between the fatty acid and the protein, retaining the fatty acid at the binding site. This hypothesis was later reinforced by the crystallographic structure of a svPLA₂-like protein denominated PrTX-II, purified from the venom of the viper carpet-jararaca (Bothrops pirajai) complexed (albeit noncovalently) with a fatty acid molecule (tridecanoic acid) bound to the active site.⁷⁷

When bound to the Lys49-PLA₂s active site, the fatty acid molecule is stabilized by a hydrogen bond with the amide group of the Cys29–Gly30 peptide bond (Figure 8).⁷⁷ Furthermore, the interaction between the carbonyl group of Cys29 and the remarkably conserved Lys122 among Lys49-PLA₂s causes this peptide bond to become hyperpolarized, further strengthening the bond with the fatty acid. In light of these findings, Lee and colleagues hypothesized that the PLA₂-like limited or absent catalytic activity could be caused by this strong interaction, which increases the affinity for the fatty acid headgroup. Thus, the product’s release from the active site would be blocked at the product release stage of the catalysis cycle, leading to enzyme inhibition.^69,71,72 In 2005, Ambrosio and colleagues proposed a synergistic activity between these enzymes.⁷¹ According to them, the PLA₂s-like proteins bind membrane fatty acids produced by the PLA₂ enzymes, stabilizing the active conformation of the PLA₂s-like proteins.^71,78

Figure 8.

Cartoon representation of the complex between Bothrops pirajai PrTX-II and a fatty acid (tridecanoic acid) (PDB 1QLL).⁷⁷ A close-up view of the binding site of PrTX-II, with the fatty acid molecule stabilized by a hydrogen bond with the Cys29–Gly30 peptide bond, which is hyperpolarized by the Lys122, increasing its affinity for the fatty acid. The Ca²⁺ binding loop interactions with the Nε atom of the Lys49 residue are also shown. PyMOL molecular graphics software package was used to generate the representations.

Based on the structural and mechanistic evidence described above, we believe that the mechanism by which PLA₂-like proteins perform phospholipid cleavage is similar to one of the svPLA₂ enzymes. Thus, an active-site water molecule, whose pK_a is lowered by hydrogen bonding to Lys49 is deprotonated by His48, which in turn is stabilized by a salt bridge to the Asp99 residue (Renetseder et al.³⁴ numbering system) and performs a nucleophilic attack on the phospholipid sn-2 bond. The positive charge of the Lys49 side chain amine plays the role of the Ca²⁺ ion by stabilizing the formal negative charge of the transition state oxyanion. Upon collapse from the transition state into the products, the lysophospholipid fragment leaves the active site. Still, the ionic hydrogen bond between the enzyme and the fatty acid carboxylate traps this reaction fragment at the binding site and interrupts the catalytic cycle.⁷⁷

5. Interfacial Activation: A Phenomenon That Increases Catalytic Activity

Secreted PLA₂ enzymes, among which are the svPLA₂, hydrolyze phospholipids in monomeric, micellar, or lipid bilayer phases.²¹ Except for group III PLA₂, these enzymes share a phenomenon termed “interfacial activation”,^17,19,21,57 which has long been captivating the scientific community.²⁹ Interfacial activation is the process by which the catalytic activity of secreted PLA₂ increases dramatically (up to 10 000-fold) when the substrate is shifted from a monomolecular dispersed form to a higher ordered and larger lipid aggregate (Figure 9).^17,19,21,57 Thus, the protein interaction with large lipid aggregates seems essential for reaching high activity in svPLA₂.^29,35 However, the phenomenon’s molecular origin is not understood despite the significant interest and effort.^19,35

Figure 9.

Representation of the svPLA (PDB 5TFV) with the phospholipid bilayer membrane, created using the CHARMM-GUI web interface.⁷⁹ The protein orientation in the membrane was automatically set up with the PPM 2.0 server, and the substrate was manually inserted. In the active center, there is a bound phospholipid substrate in which the color red represents oxygen, phosphorus is brown, nitrogen is blue, and carbon is white; the enzyme is shown in cartoon representation, and the Ca²⁺ ion is shown in dark pink. Binding to the membrane makes the enzyme more active for several orders of magnitude. PyMOL molecular graphics software package was used to generate the representations.

A related phenomenon has been observed in the triglyceride lipase from Thermomyces lanuginose.⁸⁰ When the substrate is above its solubility limit, the enzyme is noticeably more active, meaning that its reaction rate is larger toward aggregated than monodispersed substrates. The origin of this effect lies in a lipase conformational rearrangement promoted by contact with lipid aggregates.^51,80

The molecular basis of the interfacial activation of lipases inspired possibilities to explain the same phenomenon in svPLA₂ enzymes, as binding to the membrane is a fundamental step of the svPLA₂ reaction cycle. The first step of svPLA₂ hydrolysis of cell membranes or vesicles is the nonchemical adsorption to a phospholipidic surface. In this process, the enzyme buries ca. 40 phospholipids at the interfacial-adsorption surface (also called the enzyme ″i-face″); the adsorption allows substrate binding at the active site to create a Michaelis–Menten complex.^41,52 This interaction is based on electrostatic and dispersive interactions between basic residues on the PLA₂i-face and anionic phospholipids at the cell membranes or vesicles.^29,35 The i-face includes the region surrounding the hydrophobic channel connected to the active site where highly conserved residues can be found. In svPLA₂-IIA enzymes, for instance, Trp31 and Lys69 are among these residues, which have been hypothesized to have a significant role in interfacial binding. Trp31 aromatic ring may act as an anchor increasing the enzyme’s affinity for the substrate (Figure 10).^35,81

Figure 10.

Surface representation of the E. carinatus svPLA₂-IIA (PDB 1OZ6)⁸¹ hydrophobic channel, with the Trp31 and Lys69 residues in evidence. Catalytic residues are also indicated. The calcium ion is represented as an orange sphere. PyMOL molecular graphics software package was used to generate the representations.

Based on this, two models have been proposed: (1) the “substrate model” which assigns the svPLA₂ interfacial activation to the physical properties of the aggregated substrate that facilitates its diffusion through a hydrophobic channel to the catalytic site,³⁰ and (2) the “enzyme model” that suggests an enzymatic conformational change caused by the binding to the membrane interface, with the change being responsible for the enzyme’s interfacial activation.⁵¹

Concerning the “substrate model”, Scott et al.³⁰ proposed that the high lipid ordering may facilitate diffusion to the active site, reducing the conformational entropy cost of phospholipid binding. The phospholipid has many conformations in solution but far less at the active site, implying a high entropic cost for binding. In the membrane, the phospholipid conformational space is much more restricted, reducing the binding entropic penalty. In addition, the conformation of the substrate in the membrane is similar to the one at the svPLA₂ binding site. Moreover, the tight contact between the enzyme and the phospholipid surface precludes the unfavorable solvation of the tails while migrating from the membrane into the active site. Concerning the “enzyme model” even though the proposal has precedent in the lipase enzyme family, the specific conformational change, leading to a higher catalytic activity is still to be identified. Thus, the proposal lacks supporting evidence at the moment.⁵¹ An essential aspect of these proposals is that they are not mutually exclusive; both can contribute to the observed increase in reaction rate at aggregated phospholipid surfaces. Therefore, further studies are necessary to fully understand this exciting phenomenon’s molecular nature.

6. Conclusions

In this paper, several proposals for the mechanism of action of svPLA₂ are presented and critically analyzed. In light of the principles of enzyme catalysis drawn from decades of computation in similar enzymes, we propose new mechanistic possibilities that match all experimental data and have precedent in other enzymatic systems. These include a modified version of the Verheij single-water mechanism and a direct hydroxide attack mechanism. The analysis discussed in this review raises the possibility that svPLA₂s follow a range of different reaction mechanisms, depending on the volatile organization of the solvent around the active site, in particular, because the intrinsic chemistry of the several proposed mechanistic alternatives is similar. For example, the intrinsic difference in the propensity for the “single water” or the “assisted water” mechanism stems from the fine structure of the active site and the distance between the Ca²⁺ ion and His48. Given the well-known plasticity of enzymatic systems, it is expectable that different active site conformations will favor one or the other pathway. Such a scenario has been found before in aspartic proteases.⁸² Furthermore, the activity of the PLA₂ increases when the substrate is in the form of large aggregates. We review and critically analyze the “substrate model” and “enzyme model” proposals in the literature to explain this phenomenon and rationalize further the potential origin of the interfacial activation.

Over the past decades, significant advances in computational power have allowed the community to get a completer picture of the underlying enzyme catalytic features that are challenging to investigate adequately using experimental methods. Modeling enzyme-catalyzed reaction processes using hybrid quantum mechanics/molecular mechanics (QM/MM) methods has been probably the most promising approach and has been applied to the human PLA₂s in the 1990s.^83,84 Future directions for the understanding of the svPLA₂ catalytic mechanism shall pass through those methodologies, which will be decisive for discovering crucial principles underpinning the PLA₂s mechanism of action and represent a starting point for effective rational drug design.

Glossary

Abbreviations Used

α

α helix

i-face

interfacial-adsorption surface

kDa

kilodaltons

LpPLA₂s

lipoprotein-associated phospholipase A₂ enzymes

Lys49-PLA₂-like

noncatalytic phospholipase A₂

Prtx-II

piratoxin-II

QM/MM

quantum mechanics/molecular mechanics

svPLA₂

snake venom secreted phospholipase A₂

WHO

World Health Organization

Biographies

Juliana Castro-Amorim received her B.Sc. in Biology at the University of Porto and her M.Sc. in Biochemistry at the same university. Currently, she is pursuing a Ph.D. in the Laboratory of Computational Biochemistry at the University of Porto. Her research focuses on understanding the chemical mechanism of snake venom toxins and developing new treatments against snakebite envenomation.

Ana Novo de Oliveira received her Ph.D. in Pharmacy from the University of Barcelona, Spain, in 2013. She then pursued her postdoctoral studies at the University of Uppsala, Sweden. Since 2018, she has been a researcher at the LAQV-REQUIMTE center and at the University of Porto. Her main research focus is the study of biomolecular systems using a variety of theoretical and computational methods. She is particularly interested in developing new therapeutic strategies for the treatment of pathological diseases, the discovery of new drugs, understanding the structure–dynamics–function relationship in proteins, and optimizing their function through enzyme engineering.

Saulo Luís Da Silva is a Brazilian professor affiliated with the Faculty of Sciences of the University of Porto. He has a B.Sc. in Fundamental Chemistry from the University of São Paulo (Brazil, 1994) and a Ph.D. in Biochemistry from the State University of Campinas (Brazil, 2004). He has been a university professor since 1994, working in biochemistry and chemistry. His interests in investigation lie within snake venoms, toxins from animal venoms, structural biochemistry, and medicinal chemistry. He has coauthored over 80 papers and three book chapters in peer-reviewed international magazines and books and supervised ca. 40 students.

Andreimar M. Soares has a Ph.D. (2000) in Biochemistry (University of São Paulo, Brazil). He holds teaching, research, and management positions in research institutes and universities of the Brazilian Amazon region, such as the Foundation Oswaldo Cruz, Foundation for Research Support of the Rondônia State, São Lucas University, Catholic University of Rondônia, and the National Institute of Science and Technology in Epidemiology of Western Amazon. His research focuses on the isolation and functional/structural characterization of animal toxins and natural/synthetic inhibitors, identification and structural characterization of natural products of animal or plant origin with antimicrobial activity, and against neglected tropical diseases from the Amazonian and Brazilian biodiversity. He supervised more than 90 students, and published over 270 papers which have been cited more than 7000 times.

Ashis K. Mukherjee is serving as the Director of Institute of Advanced Study in Science and Technology, DST, Ministry of Science and Technology, Government of India. He received his M.Sc. (Biochemistry) from Banaras Hindu University, Varanasi (India, 1992), his Ph.D. in 1998, and his D.Sc. from Calcutta University (India, 2017). Prof. Mukherjee has been working on characterization of Indian snake venom and treatment of snakebite for more than 27 years. His current research includes proteomic analysis of snake venom, quality assessment of commercial antivenom, and cardiovascular and anticancer drug discovery from natural resources including snake venom. He has received several awards and medals for his academic performance and research achievements including Visitor’s Award for Research in Basic and Applied Sciences from the President of India in 2018.

Maria João Ramos received her Ph.D. in Quantum Chemistry from the University of Glasgow. She conducted her postdoctoral studies in Molecular Modelling at the University of Oxford (UK). She was Vice-Rector at the University of Porto, an Associate Director at the NFCR Centre for Computational Drug Discovery at Oxford, and a Director at the Science and Technology Park of the University of Porto. She is currently a Full Professor at the University of Porto, Portugal. Her research interests focus on Computational Enzymology and Drug Discovery, with a focus on snake venom chemistry and plastic biorecycling.

Pedro A. Fernandes received Ph.D. in Chemistry from the University of Porto, Portugal, in 2000. He is appointed as a Professor in the Chemistry Department of the same University. His research field is computational biochemistry, with three main lines: enzyme reaction mechanisms and enzyme engineering, structure and dynamics of proteins and membranes, and drug discovery. His most recent works focused on snake venom chemistry for developing new antidotes and transforming venom components into drugs and biorecycling plastics using enzymatic methods.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jmedchem.3c00097.

• Sequence identity matrix of six different GI/GII sPLA₂, representation of GI/GII sPLA₂ tertiary structures, a selection of available GI/GII sPLA₂ X-ray crystallographic structures in the PDB, and a sequence alignment of four classic and four PLA₂-like proteins (PDF)

The authors declare no competing financial interest.