Stabilization of adalimumab Fab through the introduction of disulfide bonds between the variable and constant domains.

Fab is a promising format for antibody drug. Therefore, efforts have been made to improve its thermal stability for therapeutic and commercial use. So far, we have attempted to introduce a disulfide bond into the Fab fragment to improve its thermal stability and demonstrated that it is possible to do this without sacrificing its biochemical function. In this study, to develop a novel stabilization strategy for Fab, we attempted to introduce a disulfide bond between the variable and constant domains and prepared three variants of Fab; H:G10C + H:P210C, L:P40C + L:E165C, and H:G10C + H:P210C + L:P40C + L:E165C. Differential scanning calorimetry measurements showed that each of these variants had improved thermal stability. In addition, the variants with two disulfide bonds demonstrated a 6.5 °C increase in their denaturation temperatures compared to wild-type Fab. The introduction of disulfide bonds was confirmed by X-ray crystallography, and the variants retained their antigen-binding activity. The variants were also found to be less aggregative than the wild type. Our results demonstrate that the introduction of a disulfide bond between the variable and constant domains significantly improves the thermal stability of Fab.

C-Terminal Cysteine PEGylation of Adalimumab Fab with an Engineered Interchain SS Bond.

Conjugation with polyethylene glycol (PEG) is performed to increase serum half-life of the Fab for clinical applications. However, current designs for recombinant Fab only allow PEGylation at the interchain SS bond (disulfide bond) at the C-terminal end of the heavy chain and light chain of the Fab, which the decrease of thermostability occurred by partial reduction of the interchain SS bond. An adalimumab Fab mutant with a novel interchain SS bond (CH1 : C177-CL : C160) and one cysteine at the C-terminal end (mutSS FabSH) was designed to maintain Fab thermostability and for site-specific PEGylation. MutSS FabSH was expressed in Pichia pastoris and purified mutSS FabSH was conjugated with 20-kDa PEG targeted at the free cysteine. Based on enzyme-linked immunosorbent assay (ELISA), PEGylation did not affect the binding capacity of the mutSS FabSH. To confirm the influence of PEGylation on the pharmacokinetic behavior of the Fab, PEGylated mutSS FabSH was administered to rats via tail vein injection. Analysis of the mean serum concentration of the PEGylated mutSS FabSH versus time through ELISA indicated an increase in half-life compared to that of non-PEGylated wild-type Fab. Consequently, we have successfully demonstrated that a Fab mutant with a novel interchain SS bond and one free cysteine at the C-terminal end can be PEGylated without changes in functionality. This design can potentially be used as a platform for modification of other recombinant Fabs.

A novel engineered interchain disulfide bond in the constant region enhances the thermostability of adalimumab Fab.

We constructed a system for expressing the Fab of the therapeutic human monoclonal antibody adalimumab at a yield of 20 mg/L in the methylotrophic yeast Pichia pastoris. To examine the contribution of interchain disulfide bonds to conformational stability, we prepared adalimumab Fab from which the interchain disulfide bond at the C-terminal region at both the CH1 and CL domains was deleted by substitution of Cys with Ala (FabΔSS). DSC measurements showed that the Tm values of FabΔSS were approximately 5 °C lower than those of wild-type Fab, suggesting that the interchain disulfide bond contributes to conformational thermostability. Using computer simulations, we designed a novel interchain disulfide bond outside the C-terminal region to increase the stability of FabΔSS. The resulting Fab (mutSS FabΔSS) had the mutations H:V177C and L:Q160C in FabΔSS, confirming the formation of the disulfide bond between CH1 and CL. The thermostability of mutSS FabΔSS was approximately 5 °C higher than that of FabΔSS. Therefore, the introduction of the designed interchain disulfide bond enhanced the thermostability of FabΔSS and mitigated the destabilization caused by partial reduction of the interchain disulfide bond at the C-terminal region, which occurs in site-specific modification such as PEGylation.

Introduction of a glycosylation site in the constant region decreases the aggregation of adalimumab Fab.

The production of therapeutic monoclonal antibodies is costly; therefore, antigen-binding fragments (Fabs) can be used instead. However, their tendency toward aggregation can reduce the half-life in the plasma and the therapeutic effectiveness. To examine the effect of glycosylation on the properties of the Fab of a therapeutic antibody, an N-glycosylation site was introduced at position 178 of the H-chain constant region of adalimumab Fab through site-directed mutagenesis of L178 N (H:L178 N Fab), and then H:L178 N Fab was expressed in Pichia pastoris. SDS-PAGE analysis with treatment of N-glycosidase F or periodic acid-Schiff reagent showed that H:L178 N Fab contained a relatively low glycan level. Moreover, the H:L178 N mutation did not decrease the binding activity and thermal stability of Fab, and H:L178 N Fab was more resistant to protease digestion than wild-type Fab. The aggregation of Fab induced by pH-shift stress was measured by monitoring the optical density at 350 nm. Although the wild-type Fab showed a large increase in optical density with an increase of protein concentration, no such increase of turbidity during aggregation was found in H:L178 N Fab. These results demonstrated that glycosylation at position 178 of the H-chain constant region of adalimumab Fab can prevent protein aggregation, and therefore serve as a potentially effective platform for drug development.

Structural characteristics and evolution of the Protobothrops elegans pancreatic phospholipase A2 gene in contrast with those of Protobothrops genus venom phospholipase A2 genes.

The nucleotide sequence of the gene encoding Protobothrops elegans (Crotalinae) pancreatic phospholipase A(2) (PLA(2)), abbreviated PePancPLA(2), was determined by means of inverted PCR techniques. Since its deduced amino acid sequence contains a pancreatic loop and shows high similarity to that of Laticauda semifasciata (Elapinae) group IB pancreatic PLA(2), PePancPLA(2) is classified into group IB PLA(2). The nucleotide sequences of the PePancPLA(2) gene, the L. semifasciata group IB pancreatic PLA(2) gene, and the L. semifasciata group IA venom PLA(2) gene are similar to one another but greatly dissimilar to those of Protobothrops genus (Crotalinae) group II venom PLA(2) genes, suggesting that the Elapinae group IB PLA(2) gene and the group IA PLA(2) gene appeared after Elapinae was established, and that the Crotalinae group II venom PLA(2) genes came into existence before Elapinae and Crotalinae diverged. A phylogenetic analysis of their amino acid sequences confirms this.

Analysis of thermostability for seven Phe to Ala and six Pro to Gly mutants in the Fab constant region of adalimumab.

To identify amino acids that play important roles in the structural stability of Fab, seven phenylalanine residues in the Fab constant region of the therapeutic antibody adalimumab were subjected to alanine mutagenesis. Six Fab mutants, H:F130A, H:F154A, H:F174A, L:F118A, L:F139A and L:F209A, showed decreased thermostability compared with wild-type Fab. In contrast, the Tm for the L:F116A mutant was 1.7°C higher than that of wild-type Fab, indicating that the F116 residue was unfavorable for Fab thermostability. Six proline mutants, H:P131G, H:P155G, H:P175G, L:P119G, L:P120G and L:P141G, were also prepared to investigate the effect of proline residues adjacent to mutated phenylalanine residues. The thermostability of the H:P155G and L:P141G mutants in particular was significantly reduced, with decreases in Tm of 5.0 and 3.0°C, respectively, compared with wild-type Fab. The H:P155 and L:P141 residues have a cis conformation, whereas the other mutated proline residues have a trans conformation. H:P155 and L:P141 had stacking interactions with the H:F154 and L:Y140, respectively, at the interface between the variable and constant regions. It is suggested that the interactions of the aromatic ring with a cis-form proline at the interface between the variable and constant regions is important for stability of Fab.

Structural characteristics and evolution of a novel venom phospholipase A2 gene from Protobothrops flavoviridis.

A novel phospholipase A(2) (PLA(2)) gene, named PfPLA 6, was found in a 6,328-bp NIS-1(5')-a segment in the Protobothrops flavoviridis (Habu, Crotalinae) genome. A comparison of the aligned nucleotide sequences of Viperidae (Viperinae and Crotalinae) venom PLA(2) genes, including PfPLA 6, revealed the deletion of a 12-bp segment called S1EX 1 and a 55-bp segment called S2EX 1 in exon 1 and the interposition of a 219-bp segment called SINT 2 (SINE) in intron 2. A classification of Viperidae PLA(2) genes based on these structural modes indicated that the A-type genes (without SINE), including PfPLA 6, are evolutionarily ancestral to the B-type (Viperinae) and C-type (Crotalinae) PLA(2) genes (both with SINE). Since PfPLA 6 is a pseudogene, an active prototype of PfPLA 6 can be assumed to be the ancestral PLA(2) gene. Putative evolutionary processes from this A-type prototype PLA(2) gene to descendent PLA(2) genes are discussed.

Effect of an intermolecular disulfide bond introduced into the first loop of CH1 domain of Adalimumab Fab on thermal stability and antigen-binding activity.

The introduction of intermolecular disulfide bonds by amino acid mutations is an effective method for stabilizing dimeric proteins. X-ray crystal structure of Fab of a therapeutic antibody, adalimumab, revealed the first loop of the CH1 domain to be partially unsolved at position 135-141. To find new sites for the introduction of intermolecular disulfide bonds in adalimumab Fab, Fab mutants targeting the unsolved region were predicted using molecular simulation software. Four Fab mutants, H:K137C-L:I117C, H:K137C-L:F209C, H:S138C-L:F116C and H:S140C-L:S114C, were expressed in the methylotrophic yeast Pichia pastoris. SDS-PAGE analysis of these mutants indicated that H:K137C-L:F209C, H:S138C-L:F116C and H:S140C-L:S114C mutants mostly formed intermolecular disulfide bonds, whereas some H:K137C-L:I117C mutants formed intermolecular disulfide bonds and some did not. Differential scanning calorimetry measurements showed increased thermal stability in all Fab mutants with engineered disulfide bonds. The bio-layer interferometry measurements, for binding of the antigen tumor necrotic factor α, indicated that Fab mutants had less antigen-binding activity than wild-type Fab. In particular, the KD value of H:K137C-L:F209C was ~17 times higher than that of wild-type Fab. Thus, we successfully introduced intermolecular disulfide bonds between the first loop region of the CH1 and CL domains and observed that it increases the thermostability of Fab and affects the antigen-binding activity.

A [Lys49]phospholipase A2 from Protobothrops flavoviridis venom induces caspase-independent apoptotic cell death accompanied by rapid plasma-membrane rupture in human leukemia cells.

Protobothrops flavoviridis venom contains plural phospholipase A(2) (PLA(2)) isozymes. A [Lys(49)]PLA(2) called BPII induced cell death in human leukemia cells. PLA2, an [Asp(49)]PLA(2) that has much stronger lipolytic activity than BPII, failed to induce cell death. BPII-treated cells showed morphological changes, DNA fragmentation, and nuclear condensation. This BPII-induced apoptotic cell death was neither inhibited by inhibitors of caspases 3 and 6 nor accompanied by activation of procaspase 3, indicating that BPII-induced cell death is caspase independent. Since inactive p-bromophenacylated BPII induced cell death, BPII-induced apoptotic cell death is independent of PLA(2) lipolytic activity. Rapid externalization of phosphatidylserine in BPII-treated cells was observed for fluorescein isothiocyanate (FITC)-labeled annexin V. In the cells treated with BPII, this spread over the cell membranes, implying that the cell toxicity of BPII is mediated via its cell-surface receptor.

Identification and evolution of venom phospholipase A2 inhibitors from Protobothrops elegans serum.

The cDNAs encoding venom phospholipase A(2) (PLA(2)) inhibitors (PLIs), named Protobothrops elegans (Pe)γPLI-A, PeγPLI-B, PeαPLI-A, and PeαPLI-B, were cloned from the P. elegans liver cDNA library. They were further divided into several constituents due to nucleotide substitutions in their open reading frames. For PeαPLI-A, two constituents, PeαPLI-A(a) and PeαPLI-A(b), were identified due to three nonsynonymous substitutions in exon 3. Far-western blot and mass-spectrometry analysis of the P. elegans serum proteins showed the presence of γPLIs, and αPLIs, which can bind venom PLA(2)s. In αPLIs from Protobothrops sera, A or B subtype-specific amino acid substitutions are concentrated only in exon 3. A comparison of γPLIs showed that γPLI-As are conserved and γPLI-Bs diversified. Mathematical analysis of the nucleotide sequences of Protobothrops γPLI-B cDNAs revealed that the particular loops in the three-finger motifs diversified by accelerated evolution. Such evolutionary features should have made serum PLIs acquire their respective inhibitory activities to adapt to venom PLA(2) isozymes.

A comprehensive analysis of novel disulfide bond introduction site into the constant domain of human Fab.

Generally, intermolecular disulfide bond contribute to the conformational protein stability. To identify sites where intermolecular disulfide bond can be introduced into the Fab's constant domain of the therapeutic IgG, Fab mutants were predicted using the MOE software, a molecular simulator, and expressed in Pichia pastoris. SDS-PAGE analysis of the prepared Fab mutants from P. pastoris indicated that among the nine analyzed Fab mutants, the F130C(H):Q124C(L), F174C(H):S176C(L), V177C(H):Q160C(L), F174C(H):S162C(L), F130C(H):S121C(L), and A145C(H):F116C(L) mutants mostly formed intermolecular disulfide bond. All these mutants showed increased thermal stability compared to that of Fab without intermolecular disulfide bond. In the other mutants, the intermolecular disulfide bond could not be completely formed, and the L132C(H):F118C(L) mutant showed only a slight decrease in binding activity and ß-helix content, owing to the exertion of adverse intermolecular disulfide bond effects. Thus, our comprehensive analysis reveals that the introduction of intermolecular disulfide bond in the Fab's constant domain is possible at various locations. These findings provide important insights for accomplishing human Fab stabilization.

Glycosylation decreases aggregation and immunogenicity of adalimumab Fab secreted from Pichia pastoris.

Glycoengineering of therapeutic proteins has been applied to improve the clinical efficacy of several therapeutics. Here, we examined the effect of glycosylation on the properties of the Fab of the therapeutic antibody, adalimumab. An N-glycosylation site was introduced at position 178 of the H chain constant region of adalimumab Fab through site-directed mutagenesis (H:L178N Fab), and the H:L178N Fab was produced in Pichia pastoris. Expressed mutant Fab contained long and short glycan chains (L-glyco Fab and S-glyco Fab, respectively). Under the condition of aggregation of Fab upon pH shift-induced stress, both of L-glyco Fab and S-glyco Fab were less prone to aggregation, with L-glyco Fab suppressing aggregation more effectively than the S-glyco Fab. Moreover, the comparison of the antigenicity of glycosylated and wild-type Fabs in mice revealed that glycosylation resulted in the suppression of antigenicity. Analysis of the pharmacokinetic behaviour of the Fab, L-glyco Fab and S-glyco Fab indicated that the half-lives of glycosylated Fabs in the rats were shorter than that of wild-type Fab, with L-glyco Fab having a shorter half-life than S-glyco Fab. Thus, we demonstrated that the glycan chain influences Fab aggregation and immunogenicity, and glycosylation reduces the elimination half-life in vivo.

Discovery of the Gene Encoding a Novel Small Serum Protein (SSP) of Protobothrops flavoviridis and the Evolution of SSPs.

Small serum proteins (SSPs) are low-molecular-weight proteins in snake serum with affinities for various venom proteins. Five SSPs, PfSSP-1 through PfSSP-5, have been reported in Protobothrops flavoviridis ("habu", Pf) serum so far. Recently, we reported that the five genes encoding these PfSSPs are arranged in tandem on a single chromosome. However, the physiological functions and evolutionary origins of the five SSPs remain poorly understood. In a detailed analysis of the habu draft genome, we found a gene encoding a novel SSP, SSP-6. Structural analysis of the genes encoding SSPs and their genomic arrangement revealed the following: (1) SSP-6 forms a third SSP subgroup; (2) SSP-5 and SSP-6 were present in all snake genomes before the divergence of non-venomous and venomous snakes, while SSP-4 was acquired only by venomous snakes; (3) the composition of paralogous SSP genes in snake genomes seems to reflect snake habitat differences; and (4) the evolutionary emergence of SSP genes is probably related to the physiological functions of SSPs, with an initial snake repertoire of SSP-6 and SSP-5. SSP-4 and its derivative, SSP-3, as well as SSP-1 and SSP-2, appear to be venom-related and were acquired later.

Unique structure (construction and configuration) and evolution of the array of small serum protein genes of Protobothrops flavoviridis snake.

The nucleotide sequence of Protobothrops flavoviridis (Pf) 30534 bp genome segment which contains genes encoding small serum proteins (SSPs) was deciphered. The genome segment contained five SSP genes (PfSSPs), PfSSP-4, PfSSP-5, PfSSP-1, PfSSP-2, and PfSSP-3 in this order and had characteristic configuration and constructions of the particular nucleotide sequences inserted. Comparison between the configurations of the inserted chicken repeat-1 (CR1) fragments of P. flavoviridis and Ophiophagus hannah (Oh) showed that the nucleotide segment encompassing from PfSSP-1 to PfSSP-2 was inverted. The inactive form of PfSSP-1, named PfSSP-1δ(Ψ), found in the intergenic region (I-Reg) between PfSSP-5 and PfSSP-1 had also been destroyed by insertions of the plural long interspersed nuclear elements (LINEs) and DNA transposons. The L2 LINE inserted into the third intron or the particular repetitive sequences inserted into the second intron structurally divided five PfSSPs into two subgroups, the Long SSP subgroup of PfSSP-1, PfSSP-2 and PfSSP-5 or the Short SSP subgroup of PfSSP-3 and PfSSP-4 The mathematical analysis also showed that PfSSPs of the Long SSP subgroup evolved alternately in an accelerated and neutral manner, whereas those of the Short SSP subgroup evolved in an accelerated manner. Moreover, the ortholog analysis of SSPs of various snakes showed that the evolutionary emerging order of SSPs was as follows: SSP-5, SSP-4, SSP-2, SSP-1, and SSP-3 The unique interpretation about accelerated evolution and the novel idea that the transposable elements such as LINEs and DNA transposons are involved in maintaining the host genome besides its own transposition natures were proposed.

SDS-induced oligomerization of Lys49-phospholipase A2 from snake venom.

Phospholipase A2 (PLA2) is one of the representative toxic components of snake venom. PLA2s are categorized into several subgroups according to the amino acid at position 49, which comprises either Asp49, Lys49, Arg49 or Ser49. Previous studies suggested that the Lys49-PLA2 assembles into an extremely stable dimer. Although the behavior on Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) under reducing or non-reducing conditions suggested the presence of intermolecular disulfide bonds, these bonds were not observed in the crystal structure of Lys49-PLA2. The reason for this discrepancy between the crystal structure and SDS-PAGE of Lys49-PLA2 remains unknown. In this study, we analyzed a Lys49-PLA2 homologue from Protobothrops flavoviridis (PflLys49-PLA2 BPII), by biophysical analyses including X-ray crystallography, SDS-PAGE, native-mass spectrometry, and analytical ultracentrifugation. The results demonstrated that PflLys49-PLA2 BPII spontaneously oligomerized in the presence of SDS, which is one of the strongest protein denaturants.

Alternative mRNA Splicing in Three Venom Families Underlying a Possible Production of Divergent Venom Proteins of the Habu Snake, Protobothrops flavoviridis.

Snake venoms are complex mixtures of toxic proteins encoded by various gene families that function synergistically to incapacitate prey. A huge repertoire of snake venom genes and proteins have been reported, and alternative splicing is suggested to be involved in the production of divergent gene transcripts. However, a genome-wide survey of the transcript repertoire and the extent of alternative splicing still remains to be determined. In this study, the comprehensive analysis of transcriptomes in the venom gland was achieved by using PacBio sequencing. Extensive alternative splicing was observed in three venom protein gene families, metalloproteinase (MP), serine protease (SP), and vascular endothelial growth factors (VEGF). Eleven MP and SP genes and a VEGF gene are expressed as a total of 81, 61, and 8 transcript variants, respectively. In the MP gene family, individual genes are transcribed into different classes of MPs by alternative splicing. We also observed trans-splicing among the clustered SP genes. No other venom genes as well as non-venom counterpart genes exhibited alternative splicing. Our results thus indicate a potential contribution of mRNA alternative and trans-splicing in the production of highly variable transcripts of venom genes in the habu snake.

The habu genome reveals accelerated evolution of venom protein genes.

Evolution of novel traits is a challenging subject in biological research. Several snake lineages developed elaborate venom systems to deliver complex protein mixtures for prey capture. To understand mechanisms involved in snake venom evolution, we decoded here the ~1.4-Gb genome of a habu, Protobothrops flavoviridis. We identified 60 snake venom protein genes (SV) and 224 non-venom paralogs (NV), belonging to 18 gene families. Molecular phylogeny reveals early divergence of SV and NV genes, suggesting that one of the four copies generated through two rounds of whole-genome duplication was modified for use as a toxin. Among them, both SV and NV genes in four major components were extensively duplicated after their diversification, but accelerated evolution is evident exclusively in the SV genes. Both venom-related SV and NV genes are significantly enriched in microchromosomes. The present study thus provides a genetic background for evolution of snake venom composition.

Discovery of novel [Arg49]phospholipase A2 isozymes from Protobothrops elegans venom and regional evolution of Crotalinae snake venom phospholipase A2 isozymes in the southwestern islands of Japan and Taiwan.

Protobothrops (formerly Trimeresurus) elegans, a Crotalinae snake, inhabits Ishigaki and Iriomote islands of the Sakishima Islands of Japan which are located between Okinawa island of Japan and Taiwan. Two phospholipase A(2) (PLA(2)) isozymes were purified to homogeneity from P. elegans venom and sequenced. This led to a discovery of novel PLA(2) isozymes with Arg at position 49, that is, [Arg(49)]PLA(2) forms, named PeBP(R)-I and PeBP(R)-II. They are polymorphic at position 3, Val for PeBP(R)-I and Ile for PeBP(R)-II. The cDNAs encoding PeBP(R)-I and PeBP(R)-II were cloned. The cDNA encoding an [Asp(49)]PLA(2) named PePLA(2) was also obtained. In contrast to PLA(2) isozymes from Protobothrops genus with 122 amino acid residues, PeBP(R)-I and PeBP(R)-II are composed of 121 amino acid residues due to lack of Pro at position 90. They exhibited necrotic and edema-inducing activities but no hemorrhagic activity was detected. A phylogenetic tree constructed for venom PLA(2) isozymes of Protobothrops genus and of related genera in the southwestern islands of Japan and Taiwan revealed that PeBP(R)-I and PeBP(R)-II of P. elegans are evolutionarily much closer to PmK49PLA(2), a [Lys(49)]PLA(2), from P. mucrosquamatus (Taiwan) than BPI and BPII, both [Lys(49)]PLA(2) forms, from P. flavoviridis (Amami-Oshima and Tokunoshima islands of Japan). Such evolutionary relationships are also seen in neutral [Asp(49)]PLA(2) isozymes from the three Protobothrops species. Thus, P. elegans is the species much closer to P. mucrosquamatus than P. flavoviridis. Their evolutionary distances seem to be well related to geological history of the islands where they have lived. In addition, it was clearly noted that Ovophis okinavensis (Amami-Oshima), which had formerly belonged to the Trimeresurus genus, and Trimeresurus stejnegeri (Taiwan) are the species fairly distant from Protobothrops genus.

Molecular diversity and accelerated evolution of C-type lectin-like proteins from snake venom.

A number of C-type lectin-like proteins that affect thrombosis and hemostasis by inhibiting or activating specific platelet membrane receptors or blood coagulation factors have been isolated from the venom of various snake species and characterized and more than 80 have been sequenced. Recent data on the primary sequences and 3D structures of C-type lectins and C-type lectin-like proteins from snake venoms have enabled us to analyze their molecular evolution. Statistical analysis of their cDNA sequences shows that C-type lectin-like proteins, with some exceptions, have evolved in an accelerated manner to acquire their diverse functions. Phylogenetic analysis shows that the A and B chains of C-type lectin-like proteins are clearly separated from C-type lectins and that the A and B chains are further divided into a group of platelet receptor-binding proteins and a group of coagulation factor-binding proteins. Elucidation of the tertiary structures of several C-type lectin-like proteins led to the discovery of a unique domain-swapping interaction between heterodimeric subunits, which creates a concave surface for ligand binding.

Amino acid sequence of a basic aspartate-49-phospholipase A2 from Trimeresurus flavoviridis venom and phylogenetic analysis of Crotalinae venom phospholipases A2.

Trimeresurus flavoviridis snakes inhabit the southwestern islands of Japan: Amami-Oshima, Tokunoshima and Okinawa. A phospholipase A2 (PLA2) of basic nature (pI 8.5) was isolated from the venom of Amami-Oshima T. flavoviridis. Its amino acid sequence determined by the ordinary procedures was completely in accord with that predicted from the nucleotide sequence of the cDNA previously cloned from Amami-Oshima T. flavoviridis venom gland, which was named PLA-B'. It consists of 122 amino acid residues and has aspartate at position 49. It induced edema in a mouse footpad assay and caused necrosis in mouse skeletal muscles. PLA-B' is similar in sequence to PLA-B (Tokunoshima) and PL-Y (Okinawa), both basic [Asp49]PLA2s, with a few amino acid substitutions, indicating occurrence of interisland mutation. Although PLA2s of Crotalinae subfamily were phylogenetically classified into four types, PLA2 (acidic or neutral [Asp49]PLA2) type, basic [Asp49]PLA2 type, neurotoxic [Asp49]PLA2 type and [Lys49]PLA2 type, it was ascertained that PLA2s of PLA2 type and [Lys49]PLA2 type are most essential as toxic components for Crotalinae snake venoms and that basic [Asp49]PLA2-type PLA2s are uniquely contained only in the venoms of T. flavoviridis species. Prediction of physiological activities of some PLA2s was made based on their location in the phylogenetic tree. Relationship of divergence of PLA2s via accelerated evolution followed by less rapid mutation and physiological activities was discussed.