Sie sind auf Seite 1von 11
Toxicon 93 (2015) 144 e 154 Contents lists available at ScienceDirect Toxicon journal homepage:

Contents lists available at ScienceDirect

Toxicon

journal homepage: www.elsevier.com/l ocate/toxicon First crotoxin-like phospholipase A 2 complex from a New

First crotoxin-like phospholipase A 2 complex from a New World non- rattlesnake species: Nigroviriditoxin, from the arboreal Neotropical snake Bothriechis nigroviridis

Bruno Lomonte a , * , Diana Mora-Obando a , Juli an Fern andez a , Libia Sanz b , Davinia Pla b , Jos e María Guti errez a , Juan J. Calvete b

b , Jos e María Guti errez a , Juan J. Calvete b a Instituto Clodomiro

a Instituto Clodomiro Picado, Facultad de Microbiología, Universidad de Costa Rica, San Jos e 11501, Costa Rica b Instituto de Biomedicina de Valencia, CSIC, Jaume Roig 11, 46010 Valencia, Spain

article info

Article history:

Received 19 September 2014 Received in revised form 21 November 2014 Accepted 27 November 2014 Available online 28 November 2014

Keywords:

Snake venom Viperidae Phospholipase A 2 Crotoxin Bothriechis nigroviridis

abstract

Bothriechis nigroviridis is an arboreal Neotropical pitviper found in Costa Rica and Panam a. A previous proteomic pro ling of its venom revealed the presence of proteins with homology to the A and B sub- units of crotoxin/Mojave toxin, a heterodimeric phospholipase A 2 (PLA 2 ) complex only described in rattlesnake venoms (genera Crotalus and Sistrurus ). The native crotoxin-like heterodimer, named nigroviriditoxin, and its A and B subunits were isolated in the present work, and the complete amino acid sequence of the B subunit was determined. The puri ed A and B components were demonstrated to form a complex when reconstituted under native conditions. Nigroviriditoxin presents features similar to crotoxin, albeit displaying lower toxicity: the A component decreases the PLA 2 activity of the B component, and increases its lethal potency in mice. Also in similarity to crotoxin B, nigroviriditoxin B induces myonecrosis. Its 122 amino acid sequence presents 81% identity with crotoxin B. Accordingly, nigroviriditoxin B was cross-recognized by equine antibodies from a Crotalus durissus terri cus anti- venom. Phylogenetic analysis shows that the novel PLA 2 from B. nigroviridis venom is basal to the branch including all the homologous PLA 2 enzymes described in rattlesnakes, and more distant from PLA 2 s from Bothriechis species. Nigroviriditoxin is the rst heterodimeric PLA 2 complex found in a non-rattlesnake, Neotropical viperid venom, which displays structural, functional, and immunochemical similarities to crotoxin. The present ndings are compatible with the existence of the particular structural trait of crotoxin-like molecules in New World pitvipers before the split of the Meso-South American and the Nearctic clades.

© 2014 Elsevier Ltd. All rights reserved.

1. Introduction

The black-speckled palm snake, Bothriechis nigroviridis , is an arboreal Neotropical pit viper that inhabits subtropical rainforests and temperate forests at medium to high elevations (1150 e 3000 m) on the Cordillera de Tilar an (highlands of Mon- teverde) and Cordillera Volc anica Central in Costa Rica, southeast- ward through the Cordillera de Talamanca to the Chiriquí province in Panam a ( Savage, 2002; Campbell and Lamar, 2004; Sol orzano, 2004 ). Its Latin name originates from its distinctive speckled co- lor pattern combining black ( nigro ) and emerald green ( viridis )

* Corresponding author. E-mail address: bruno.lomonte@ucr.ac.cr (B. Lomonte).

0041-0101/ © 2014 Elsevier Ltd. All rights reserved.

spots. Adults have been described to prey on small rodents, lizards, frogs, and occasionally small birds ( Sol orzano, 2004 ). Owing to its con nement to undisturbed habitats with little human contact, envenomings by this pitviper species appear to be extremely un- common, as no documented cases could be found in a literature search. In a previous study, the venom of B. nigroviridis was analyzed by a combined proteomic and toxicological approach ( Fern andez et al., 2010 ), revealing remarkable differences with the venoms of three other Bothriechis species found in Costa Rica: Bothriechis lateralis , Bothriechis schlegelii ( Lomonte et al., 2008 ) and Bothriechis supra- ciliaris ( Lomonte et al., 2012 ). Surprisingly, B. nigroviridis venom is devoid of metalloproteinases ( Fern andez et al., 2010 ), a widespread and often predominant protein family in viperid venoms ( Calvete et al., 2007; Lomonte et al., 2014 ). In agreement with its lack of

B. Lomonte et al. / Toxicon 93 (2015) 144 e154

145

metalloproteinases, B. nigroviridis venom does not induce hemor- rhage in the mouse skin assay ( Fern andez et al., 2010 ). Another striking feature revealed by the proteomic study of this venom was the identi cation of a major phospholipase A 2 (PLA 2 ; EC 3.1.1.4) component which showed signi cant amino acid sequence ho- mology to crotoxin B from rattlesnakes ( Fern andez et al., 2010 ). In addition, this venom presented a protein with homology with the acidic subunit of crotoxin, i.e. crotoxin A ( Fern andez et al., 2010 ). Crotoxin is a well characterized heterodimeric PLA 2 complex ( Aird et al., 1986; Faure et al., 1994; Marchi-Salvador et al., 2008; Sampaio et al., 2010 ) thus far only found in the venom of rattlesnakes (genera Crotalus and Sistrurus ). The unusual presence of crotoxin- like components outside rattlesnake venoms prompted us to isolate the native crotoxin-like complex, and characterize its functional properties, the amino acid sequence of the PLA 2 component, and its phylogenetic relationships to homologous proteins found in other crotalid snake venoms.

2.4. SDS-PAGE and nESI-mass spectrometry

Peaks corresponding to the A and B subunits of nigroviriditoxin obtained from RP-HPLC were redissolved in water, and their protein concentration was estimated by measuring the absorbance at

280 nm in a NanoDrop 2000c instrument (Thermo Scienti c).

Samples were then analyzed by SDS-PAGE using pre-cast gradient gels (4 e 20%; Bio-Rad) under reducing or non-reducing conditions.

Proteins were visualized by Coomassie blue R-250 staining, and recorded using a ChemiDoc gel imager with ImageLab software (Bio-Rad). Molecular mass markers were run in parallel. The isotope-averaged molecular mass of puri ed A and B subunits, respectively, was determined by nESI-MS on a QTrap-3200 mass spectrometer (Applied Biosystems). The proteins were diluted in 50% acetonitrile containing 0.1% formic acid, and directly infused into a nanospray source for ionization at 1200 V. Analysis was performed in positive enhanced multicharge mode in the range 600 e1700 m / z , aided by the Analyst v.1.5 software (ABSciex).

2. Materials and methods

2.1. Venom

Venom was collected and pooled from three adult specimens of B. nigroviridis kept at the serpentarium of Instituto Clodomiro Pic- ado, University of Costa Rica. After centrifugation to remove insoluble matter, the venom was lyophilized and stored at 20 C until use.

2.5. Complex formation between nigroviriditoxin A and B subunits

The A and B subunits of nigroviriditoxin were incubated together or alone for 15 min at room temperature, and then sub- jected to agarose gel electrophoresis under native conditions. The agarose was dissolved at 1% (w/v) in 0.116 M Tris, 0.3 M glycine, pH 8.6 buffer. Electrophoresis was performed at 2 mA/cm for 90 min, and proteins were visualized by Coomassie R-250 staining.

2.6. Amino acid sequencing of nigroviriditoxin B

2.2. Isolation and characterization of native nigroviriditoxin

heterodimer

To demonstrate the existence of the nigroviriditoxin AB heter- odimer, 50 mg of whole venom in 100 mL of 0.1 M ammonium ac- etate, pH 6.9, were fractionated using an ETTAN-LC chromatograph and a Bio-Sil SEC 250 (300 7.8 mm) size-exclusion chromato- graphic column (Bio-Rad) equilibrated in the same buffer. The eluate was monitored at 215 nm and the fractions, collected manually, were analyzed by LC-MS using a nano-Acquity Ultra- Performance LC ® (UPLC ® ) using a BEH130 C 18 (100 mm 100 mm,

1.7 mm particle size) column in-line with a Waters SYNAPT G2 High

De nition Mass Spectrometry System. The ow rate was set to

0.6 ml/min and the column was developed with a linear gradient of

0.1% formic acid in water (solution A) and 0.1% formic acid in acetonitrile (solution B), isocratically 1% B for 1 min, followed by 1 e12% B for 1 min, 12 e 40% B for 15 min, and 40 e 85% B for 2 min.

2.3. Isolation of nigroviriditoxin A and B subunits

Venom aliquots of 2.0 e 2.5 mg were dissolved in 220 mL of so- lution A (0.1% tri uoroacetic acid in water) and centrifuged at 3000 g for 5 min. The clear supernatant was fractionated by reverse-phase HPLC, using an Agilent Model 1200 chromatograph. Two-hundred microliters of venom were injected to a C 18 column (Teknochroma; 250 4.6 mm; 5 mm particle size) equilibrated with solution A, and elution was then carried out at a ow rate of 1 mL/ min by applying the following gradient toward solution B (0.1% TFA in acetonitrile): 0% B for 5 min, 0 e15% B over 10 min, 15 e 45% B over 60 min, 45 e70% B over 10 min, and 70% B over 9 min. The eluant was monitored at 215 nm, and the fractions of interest were collected manually, dried by vacuum centrifugation at 45 C, and stored at 20 C.

The N-terminal amino acid sequence of nigroviriditoxin B was obtained by automated direct Edman degradation in a Procise in- strument (Applied Biosystems) following manufacturer's in- structions ( Fern andez et al., 2010 ). The protein sequence was completed by tandem mass spectrometry of peptides obtained af- ter the digestion of the DTT-reduced and iodoacetamide-alkylated protein with trypsin or chymotrypsin. Peptides were diluted 1:2

with a saturated a-cyanohydroxycinnamic acid solution in 50% acetonitrile and 0.1% tri uoroacetic acid, spotted onto an Opti-TOF-

384 plate, dried, and analyzed by MALDI-TOF-TOF on a Model

4800-Plus Proteomics Analyzer (Applied Biosystems). Spectra were acquired in positive re ector mode at 2 kV in the 875 e 4000 m / z range using a laser intensity of 3000 and 1500 shots ( Lomonte et al., 2012 ). CalMix standards (ABSciex) spotted on the same plate were used for external calibration. Few peptides were analyzed by nESI- MS/MS on a QTrap2000 mass spectrometer (Applied Biosystems) by direct infusion into a nanospray source. Selected doubly- or triply-charged ions from spectra obtained in enhanced resolution mode (250 amu/s) were subjected to fragmentation using the enhanced product ion option with Q 0 trapping. Settings were: Q1, unit resolution; collision energy, 25 e 40 eV; linear ion trap Q3 ll time, 250 ms; and Q3 scan rate, 1000 amu/s ( Calvete et al., 2007). All fragmentation spectra obtained by MALDI- or ESI-mass spec- trometry were interpreted manually to derive de novo amino acid sequences.

2.7. Phospholipase A 2 activity

The PLA 2 activity of nigroviriditoxin B was determined on the monodisperse synthetic substrate 4-nitro-3-octanoyl-benzoic acid (NOBA) ( Holzer and Mackessy, 1996 ). Various amounts of the toxin, dissolved in 25 mL of 10 mM Tris, 10 mM CaCl 2 , 0.1 M NaCl, pH 8.0 buffer, were added to 200 mL of this buffer in triplicate wells of a microplate. After mixing, 25 mL of NOBA (1 mg/mL in acetonitrile) were added, to achieve a nal substrate concentration of 0.32 mM.

146

B. Lomonte et al. / Toxicon 93 (2015) 144 e154

The mixtures were incubated for 60 min at 37 C, and absorbances at 405 nm were recorded. PLA 2 activity was expressed as the nal change in absorbance 1000. For comparison, puri ed B subunit of

crotoxin from Crotalus durissus terri cus (kindly provided by Prof.

^

Andreimar Soares, Universidade Federal de Rond onia, Brazil) was included in the assay. In another experiment, the effect of the A component of nigroviriditoxin on the PLA 2 activity of the B component was tested by mixing both at 1.5:1 or 3:1 M ratios (A:B) and then comparing enzyme activity to that of component B alone, as described above.

2.8. Myotoxic activity

The ability of nigroviriditoxin B to induce skeletal muscle ne- crosis was evaluated in groups of ve CD-1 mice of 18 e 20 g of body weight, following protocols approved by the Institutional Com- mittee for the Use and Care of Animals (CICUA) at University of Costa Rica. Twenty mg of the toxin were injected intramuscularly in the gastrocnemius, in a volume of 100 mL of PBS (0.12 M NaCl, 0.04 M sodium phosphate buffer, pH 7.2). A control group received an identical injection of PBS alone. After 3 h, a blood sample was obtained from the tail of each animal into a heparinized capillary tube, and after centrifugation, 4 mL of plasma were used to deter- mine the creatine kinase (CK) activity using a UV-kinetic assay (CK- Nac, Biocon Diagnostik). CK activity was expressed in units/L. For comparison, 20 mg of crotoxin B was injected into another group of mice and the plasma CK activity was assessed as described.

2.9. Lethal activity of nigroviriditoxin B and potentiation by

nigroviriditoxin A

Variable doses of nigroviriditoxin B (10 e 80 mg) were injected intravenously (i.v.) in the tail vein in groups of four mice of 16 e18 g of body weight, in 100 mL of PBS. Deaths were recorded after 48 h and the median lethal dose (LD 50 ± S.E.) was calculated by probits ( Finney, 1971 ) using the BioStat v.2009 software (AnalystSoft). A similar assay was performed for the isolated subunit A component, using doses up to 70 mg. For evaluation of synergy, the natural proportion of the A and B components of nigroviriditoxin in the venom was estimated by integrating their corresponding chro- matographic peak areas in the RP-HPLC signal at 215 nm, roughly corresponding to peptide bond abundance. The molar proportion of A:B components was estimated to be approximately 1:2. On this basis, the lethality of mixtures of both components was tested in mice at 1:2 M ratio, and compared to the results obtained when each component was assayed independently.

2.10. Immunochemical cross-recognition of nigroviriditoxin B by antivenom against C. d. terri cus

Puri ed nigroviriditoxin A subunit, B subunit, or crotoxin B, respectively, were adsorbed onto wells of an ELISA plate (Nunc Maxisorp) by incubating 0.2 mg/well of the proteins in 100 mL of 0.1 M Tris, 0.15 M NaCl, pH 9.0 buffer overnight at 4 C. After washing ve times with PBS, free sites were blocked with PBS containing 1% bovine serum albumin (BSA) for 1 h. Then, serial dilutions of equine antivenom against C. d. terri cus venom (Insti- tuto Butantan, batch 0309-38), or normal equine serum as a negative control, were added and incubated 1 h. Plates were then washed with 0.05 M Tris, 0.15 M NaCl, 20 mM ZnCl 2 , 1 mM MgCl 2 (pH 7.4), and bound antibodies were detected by incubation with an anti-horse total IgG-alkaline phosphatase conjugate (1:3000; Sigma) for 1 h, followed by washing and color development with p - nitrophenylphosphate (1 mg/mL) in diethanolamine buffer (pH

9.8). Absorbances were recorded on a microplate reader (Labsys- tems) at 405 nm. Assays were performed in triplicate wells.

2.11. Phylogenetic analysis of nigroviriditoxin B

Proteins showing amino acid sequence identity values of at least 75% and a score of at least 191 in comparison to nigroviriditoxin B were retrieved after a BLAST search ( http://blast.ncbi.nlm.gov ). These proteins, together with three available sequences corre- sponding to Asp49-PLA 2 s characterized from the genus Bothriechis (Q6EER4, C0HJC1, A8E2V4), were aligned with the program MUS- CLE ( Edgar, 2004 ) using the MEGA6 software ( Tamura et al., 2013 ). The evolutionary history was inferred with MEGA6, by using the maximum likelihood method based on the JTT matrix-based model ( Jones et al., 1992 ). Three sequences corresponding to PLA 2 s from elapid snake species (P81167 from Micrurus nigrocinctus , P00605 from Naja nigricollis and P00614 from Oxyuranus scutellatus scu- tellatus ) were used as an outgroup in this analysis, which involved a total of 22 amino acid sequences.

2.12. Homology modeling of nigroviriditoxin B

The Swiss-Model automated homology modeler ( http:// swissmodel.expasy.org/ ; Biasini et al., 2014; Arnold et al., 2006 ) was used to predict the three-dimensional structure of nigrovir- iditoxin B. The modeler selected crotoxin B (PDB access code 3R0L; Faure et al., 2011 ) as the best template, with a resolution of 1.35 Å and 80.5% sequence identity with the target. Swiss-PDB viewer v.4.1 was used to superimpose the obtained model to the template, and to calculate r.m.s.d values for main chain a-carbons and back- bones. Electrostatic surface potential of the proteins were repre- sented with the DS ViewerPro v.6.0 software (Accelrys).

3. Results

Fractionation of B. nigroviridis venom by RP-HPLC on C 18 is shown in Fig. 1. The peaks eluting at ~42 and ~51 min had been previously identi ed to have homology to the acidic (A) and basic (B) subunits of crotoxin, respectively, in a proteomic study on this venom ( Fern andez et al., 2010 ). Therefore, we explored if these two proteins could form a crotoxin-like complex, hereby named nigroviriditoxin. By SDS-PAGE, the PLA 2 B subunit migrated as a band of ~15 kDa under reducing conditions, whereas the A-chain migrated as a main band of ~9 kDa ( Fig. 2 ). Under non-reducing conditions, the A chain was observed at ~15 kDa, suggesting a possible dimerization or an anomalous migration. Similarly, the B chain appeared to aggregate, as evidenced by a continuous smear in the 15 e 25 kDa range ( Fig. 2 ), commonly observed under non-

e 25 kDa range ( Fig. 2 ), commonly observed under non- Fig. 1. Isolation of

Fig. 1. Isolation of the A and B subunits of nigroviriditoxin from the venom of Bothriechis nigroviridis . Crude venom (~2 mg) was fractionated by RP-HPLC on a C 18 column (4.6 250 mm) and monitored at 215 nm. Protein elution was performed with an acetonitrile gradient (dashed line) as described in Materials and methods . The peaks corresponding to A (acidic chain) and B (PLA 2 ) subunits are labeled.

B. Lomonte et al. / Toxicon 93 (2015) 144 e154

147

B. Lomonte et al. / Toxicon 93 (2015) 144 e 154 147 Fig. 2. Electrophoretic mobility

Fig. 2. Electrophoretic mobility of the A and B subunits of nigroviriditoxin by SDS- PAGE (4 e20% gradient gel). Samples were analyzed after reducing (r ) or non- reducing (nr) conditions) and visualized by Coomassie blue R-250 staining. Values for the molecular mass markers (m) are shown to the right, in kDa. (For interpretation of the references to color in this gure legend, the reader is referred to the web version of this article.)

reducing conditions in group II PLA 2 s of viperid snake venoms

( Soares et al., 2000; Angulo et al., 2000; Nú nez et al., 2004 ),

~

including crotoxin ( Faure et al., 1994 ). By ESI-MS analysis, the A subunit presented an isotope-averaged molecular mass of 9607 Da ( ± 1), whereas the B subunit showed a main molecular mass of 14,083 Da, and an additional mass of 14,113 Da ( ± 2) ( Fig. 3 ). In the latter case, the observed mass heterogeneity likely correspond to the presence of PLA 2 isoforms, a frequent nding in snake venoms ( Doley et al., 2010 ) and in crotoxin preparations, due to the complex multigene nature of this protein family ( Faure et al., 1994; Faure and Bon, 1988 ). We explored the capability of the two proteins to associate into a crotoxin-like complex using two different approaches. On the one hand, B. nigroviridis venom was fractionated by size-exclusion chromatography and the composition of the resulting fractions was analyzed by mass spectrometry. A major peak eluted at 10.1 mL ( Fig. 4 A), corresponding to the position expected for a 22.9 kDa standard protein ( Fig. 4 B), and contained components of 9605.6 Da (A 1 ), 9421.5 Da (A 2 ), and 14,081.3 Da (B) ( Fig. 4 C, upper panel). Reverse-phase chromatography prior to MS analysis separated the 14 kDa ( Fig. 4 C, middle panel) from the 9 kDa ( Fig. 4 C, lower panel) proteins. As a whole, these data clearly indicate that the crotoxin- like proteins of B. nigroviridis associate into heterodimers, hereby named nigroviriditoxin A 1 B and A 2 B. The capability of the RP-HPLC-dissociated and isolated A and B components of nigroviriditoxin to reconstitute the heterodimer(s) was also analyzed by agarose electrophoresis under native condi- tions. The free A component of nigroviriditoxin migrated rapidly

towards the anode, indicating its acidic character, whereas the B

the anode, indicating its acidic character, whereas the B Fig. 3. Molecular mass of the A

Fig. 3. Molecular mass of the A and B subunits of nigroviriditoxin as determined by nano-electrospray ionization mass spectrometry (ESI-MS). The multiply-charged ion series for the A (A) and B (C) subunits are deconvoluted in panels (B) and (D) , respectively.

148

B. Lomonte et al. / Toxicon 93 (2015) 144 e154

148 B. Lomonte et al. / Toxicon 93 (2015) 144 e 154 Fig. 4. Panel A,

Fig. 4. Panel A, fractionation of B. nigroviridis venom (green trace) by size-exclusion chromatography. The column was calibrated with a mixture of dextran blue (2000 kDa), bovine serum albumin (66.4 kDa), equine cytochrome c (12.3 kDa), and vitamin B12 (1.35 kDa), and used for estimating the apparent molecular mass of nigroviriditoxin (peak labeled with an asterisk) as 22.9 kDa (panel B). The heterodimeric association of nigroviriditoxin A and B subunits was demonstrated by mass spectrometry (panel C). The major peak eluting at 10.1 mL contained components of 9605.6 Da (A 1 ), 9421.5 Da (A 2 ) and 14,081.3 Da (B) (upper panel). Reverse-phase chromatographic separation prior to MS analysis separated the 14 kDa B-subunit (middle panel) from the 9 kDa A1 and A2 subunits (lower panel). (For interpretation of the references to color in this gure legend, the reader is referred to the web version of this article.)

B. Lomonte et al. / Toxicon 93 (2015) 144 e154

149

B. Lomonte et al. / Toxicon 93 (2015) 144 e 154 149 Fig. 5. Complex formation

Fig. 5. Complex formation between the A and B subunits of nigroviriditoxin evaluated by native agarose gel electrophoresis at pH 8.6. Lane 1 : A subunit; lane 2: B subunit; lane 3 : mixture of A and B subunits, at 1.5:1 (A:B) molar ratio; lane 4: crotoxin B; lane 5: mixture of nigroviriditoxin A subunit and crotoxin B, at 1.5:1 (A:B) molar ratio. Anode ( þ ) and cathode ( ) positions are indicated. Proteins were loaded at 6 mg/lane and visualized by Coomassie R-250 staining after electrophoresis.

component migrated toward the cathode, indicating its basic pI ( Fig. 5 ). When these two components were mixed, a new band appeared at half-way the migration distance between both free

components ( Fig. 5 ), clearly indicating the formation of an A þ B complex. Accordingly, an evident reduction in the intensity of the A subunit was observed, and no free B subunit was detected. Further evidence for the similarity of the A component of nigroviriditoxin to crotoxin A was obtained in this native gel assay, by observing that

a mixture of the former with crotoxin B also results in the formation of a complex, concomitant with the disappearance of free crotoxin

B ( Fig. 5 ).

Due to the limited availability of B. nigroviridis venom, which yielded only small amounts of isolated nigroviriditoxin A ( Fig. 1 ), and considering its expected structural complexity (assuming similarity to crotoxin A, formed by three covalently-linked chains generated by proteolytic processing at the same polypeptide bonds as crotoxin; Fern andez et al., 2010 ), the determination of the amino acid sequence of this component could not be attempted. On the other hand, the complete sequence of nigroviriditoxin B was determined by the combination of N-terminal Edman degradation and tandem MS of proteolytic peptides ( Fig. 6 ). Only the C-terminal

cysteine residue of this sequence could not be directly observed in the tryptic peptide ion of m / z 946.1 þ3 due to the preceding lysine, but was inferred on the basis of its absolute conservation in group II PLA 2 s from snake venoms ( Arni and Ward, 1996 ), and the isotope- averaged molecular mass determined for the native protein. The sequence of nigroviriditoxin B consists of 122 amino acid residues, with all canonical cysteine positions of group II PLA 2 s being conserved. This sequence will appear in UniProt ( http://www. uniprot.org/ ) under the accession number C0HJL8. The theoretical molecular mass predicted by the sequence is 14,112 Da, which presents a difference of ~29 e 30 Da with the most abundant form (14,083 Da) or coincides, within instrumental error, with the sec- ond form observed (14,113 Da) by ESI-MS analysis ( Fig. 3 ). Ac- cording to its sequence, the theoretically expected pI of this protein would be 8.5, in agreement with its observed cathodic migration in the native agarose gel system ( Fig. 5 ). Nigroviriditoxin B presents sequence homology to crotoxin B isoforms described in the venom of C. d. terri cus ( Fig. 7 ), with identity values ranging from 77 to 80%. The highest identity (81%) was observed in comparison to a PLA 2 from Sistrurus catenatus tergeminus (Q6EER2; Chen et al., 2004 ). By multiple alignment of nigroviriditoxin B with PLA 2 s having 75% identity, and with three Asp49-PLA 2 s described in the genus Bothriechis, a phylogenetic tree was inferred ( Fig. 8 ). Elapid PLA 2 s served as an outgroup, as they belong to the group I classi cation of this enzyme superfamily. Noteworthy, nigroviriditoxin B appears basal to the group con- taining all the different crotoxin B variants described in rattle- snakes, i.e. Crotalus and Sistrurus species, and does not cluster together with other PLA 2 s characterized from venoms of species of the genus Bothriechis ( Fig. 8 ). The branch of nigroviriditoxin B stems from acidic PLA 2 s of Old-world crotalid species ( Gloydius halys , Ovophis monticola ) and three acidic PLA 2 s isolated from B. lateralis and B. schlegelii ( Fig. 8 ). The homology of nigroviriditoxin B to the various crotoxin B isoforms found in the venoms of rattlesnakes is in agreement with its functional characteristics. As shown in Fig. 9 A, isolated nigro- viriditoxin B displays functional PLA 2 activity, only slightly lower in comparison to crotoxin B. Also in similarity with the latter, nigro- viriditoxin B induced myonecrosis when injected intramuscularly in mice, although this effect was markedly lower in comparison to crotoxin B ( Fig. 9 B). Regarding lethal activity, nigroviriditoxin B alone showed an LD 50 of 50 (44.9 e 55.6) mg/mouse (i.e., 2.9 mg/g body weight) when injected intravenously. Mice showed signs of respiratory paralysis before death. When nigroviriditoxin A was added to the B component, at 1:2 (A:B) molar ratio, the potency of

to the B component, at 1:2 (A:B) molar ratio, the potency of Fig. 6. Amino acid

Fig. 6. Amino acid sequence of nigroviriditoxin B. Peptides generated by the digestion of the protein with trypsin (T) or chymotrypsin (C) were de novo sequenced by mass spectrometry. The rst 60 amino acid residues were determined by N-terminal Edman degradation.

150

B. Lomonte et al. / Toxicon 93 (2015) 144 e154

150 B. Lomonte et al. / Toxicon 93 (2015) 144 e 154 Fig. 7. Multiple alignment

Fig. 7. Multiple alignment of the amino acid sequence of nigroviriditoxin B with crotalid venom phospholipases A 2 selected on the basis of sequence identity values of at least 75% and a BLAST score of at least 191. Three available sequences of phospholipases A 2 from Bothriechis venoms were additionally included. Sequences were aligned using MUSCLE (Edgar, 2004) with the MEGA6 software ( Tamura et al., 2013 ), as described in Materials and methods . Protein access codes correspond to the UniProtKB database at the ExPASy Proteomics Server. Identical amino acid positions are shaded in gray, with cysteine residues in boldface. Isoelectric points (pI) were calculated with the compute pI/MW tool ( http://web.expasy. org/compute_pi/), and the basic or acidic proteins are represented in blue or red colors, respectively. (For interpretation of the references to color in this gure legend, the reader is referred to the web version of this article.)

this effect increased by 60%, to an LD 50 of 31 (27.5 e 33.9) mg/mouse (2.2 mg/g body weight). Higher amounts of A subunit could not be tested due to limitations in venom availability. The isolated A subunit was not lethal per se , and did not cause any evident alter- ations, when injected at doses up to 70 mg/mouse (4.1 mg/g body weight). The addition of the acidic A component to nigroviriditoxin B caused a signi cant reduction of its PLA 2 activity in vitro , nearly by half ( Fig. 10 ). On the other hand, in ELISA experiments nigrovir- iditoxin B was clearly immunorecognized by an equine antivenom raised against the venom of C. d. terri cus ( Fig. 11 A), with a signal nearly as intense as that generated by puri ed crotoxin B ( Fig. 11 C). Recognition of nigroviriditoxin A by this antivenom was negligible ( Fig. 11 B), although in this case the unavailability of crotoxin A did not allow to control for the level of antibodies against the homol- ogous antigen in the antivenom. The three-dimensional structure of nigroviriditoxin B was modeled using the crystal structure of crotoxin B as a template. The obtained model appears virtually superimposable to the latter structure, with r.m.s.d. values for a-carbon and backbone atoms of 0.06 Å and 0.08 Å, respectively ( Fig. 12 ). Both proteins differ at 24 out of their 122 amino acid residues, and the corresponding side

chains of these are represented in Fig. 12 B. Out of the 22 amino acid residues of crotoxin B proposed to interact with crotoxin A in the crystallized complex ( Faure et al., 2011 ), nigroviriditoxin B con- serves 18. The only four amino acid residues that change from crotoxin B to nigroviriditoxin B in this particular set include His 1 / Asn, Lys 14 /Arg, Trp 61 /Ser, and Pro 111 /Leu. The high structural simi- larity between nigroviriditoxin B and crotoxin B is also re ected by their conserved surface charge distributions, compared in Fig. 12 C and D).

4. Discussion

A previous proteomic study on the venom of B. nigroviridis revealed several unusual characteristics ( Fern andez et al., 2010 ). Among them, the unexpected occurrence of a predominant PLA 2 with similarity to crotoxin B, together with a protein resembling the acidic A subunit of crotoxin, prompted us to investigate whether this venom would contain a crotoxin-like complex, to date only found in the venoms of rattlesnakes ( Crotalus and Sistrurus ) among New World pitvipers. Crotoxin, rst isolated from the South American rattlesnake C. d. terri cus , is the major toxic component in the venom of several rattlesnake species. It is a heterodimeric

B. Lomonte et al. / Toxicon 93 (2015) 144 e154

151

B. Lomonte et al. / Toxicon 93 (2015) 144 e 154 151 Fig. 8. Inferred phylogenetic

Fig. 8. Inferred phylogenetic relationships of nigroviriditoxin B. The amino acid sequences aligned in Fig. 7, together with three sequences corresponding to phospholipases A 2 from elapids used as an outgroup (P81167 from Micrurus nigrocinctus, P00605 from Naja nigricollis and P00614 from Oxyuranus s. scutellatus ), were used to construct a tree using the maximum likelihood method based on the JTT matrix-based model, implemented in MEGA6. The tree is drawn to scale, with branch lengths measured in the number of sub- stitutions per site. Codes for basic and acidic proteins are represented in blue and red, respectively. (For interpretation of the references to color in this gure legend, the reader is referred to the web version of this article.)

complex composed of a basic PLA 2 subunit (crotoxin B) and an acidic, non-enzymatic subunit (crotoxin A, crotapotin), held together non-covalently ( Sampaio et al., 2010 ). A similar complex found in various North American rattlesnakes such as Crotalus scutulatus scutulatus has been named Mojave toxin ( Hendon, 1975 ). Although PLA 2 s with homology to crotoxin B have been found in non-rattlesnake venoms (for example, agkistrodotoxin; Chen et al., 2004 ), none have been shown to form a complex with a subunit A- like component from the same venom. However, reconstitution experiments of agkistrodotoxin from Agkistrodon blomhof brevi- caudus with crotoxin A from C. d. terri cus , have shown that this potential exists ( Choumet et al., 1993 ). Thus, the term crotoxin- like has been used until now to indicate amino acid sequence similarity between a given PLA 2 and crotoxin B ( Chen et al., 2004 ), but not similarity regarding the formation of its characteristic A þ B heterodimeric complex. This study con rms that B. nigroviridis venom contains a het- erodimeric PLA 2 built by A and B components with homology to crotoxin A and B subunits. In addition, our work demonstrates the ability of the isolated A and B subunits to reconstitute the complex under native conditions. The ability of the nigroviriditoxin A component to form an inter-species complex with crotoxin B as well, strongly suggests a high degree of conservation of the struc- tural determinants needed for subunit interaction in these two related heterodimeric toxins. The low yield of subunit A, and its likely multichain structure, precluded attempts to determine its amino acid sequence. On the other hand, the complete sequence of nigroviriditoxin B was determined. Identity values from 75 to 81% were observed by its comparison with basic PLA 2 s from Crotalus and Sistrurus venoms, although at least two acidic PLA 2 s from the Asian genera Ovophis and Gloydius showed sequence identities of 75%. Interestingly, three available Asp49-PLA 2 sequences from other Bothriechis

venoms showed lower identity (52 e 68%) with nigroviriditoxin B, than PLA 2 s from rattlesnakes. The phylogenetic tree inferred by these sequence comparisons positioned nigroviriditoxin B basal to the branch that includes all members of the group of PLA 2 s from rattlesnakes. Consistent with hypotheses on the invasion of viperid ancestors from the Old World into North America ( Van Devender and Conant, 1990; Douglas et al., 2006; Castoe and Parkinson, 2006; Castoe et al., 2009 ), the two PLA 2 s from Asian genera are basal to nigroviriditoxin B in the tree, which is in concordance with the proposal of Old World genera ( Protobothrops , Ovophis , Trimer- esurus , and Gloydius ) as potential sister taxa to the New World clade ( Malhotra and Thorpe, 2004; Wüster et al., 2008 ). The closer phylogenetic relationship of nigroviriditoxin B to PLA 2 s from rat- tlesnakes, and its separation from other PLA 2 s found in Bothriechis species to date, suggest the divergence of these enzymes before the diversication of this genus across Middle America, estimated to have occurred 12 e16 Mya ( Castoe et al., 2009 ). Our ndings are compatible with the existence of the particular structural trait of crotoxin-like molecules in New World pitvipers, prior to the split of the Meso-South American and the Neartic clades ( Castoe and Parkinson, 2006 ). Evident structural similarities between nigroviriditoxin B and crotoxin B are additionally supported at the functional level. Nigroviriditoxin B displays PLA 2 activity in vitro , only slightly lower than that of crotoxin B. Although weaker than crotoxin B, nigro- viriditoxin B is also myotoxic in vivo. Furthermore, as observed for crotoxin ( Faure et al., 1993 ), the acidic A component of nigrovir- iditoxin modulates both the enzymatic and toxic activities of the basic PLA 2 subunit, albeit the effect of complex formation on these two activities appears to be less pronounced than in the case of crotoxin. The lethal potency of nigroviriditoxin B by the i.v. route is weaker than that of crotoxin B, whose i.v. LD 50 has been reported at 0.7 mg/g body weight ( Faure et al., 1993 ). Altogether, these results

152

B. Lomonte et al. / Toxicon 93 (2015) 144 e154

152 B. Lomonte et al. / Toxicon 93 (2015) 144 e 154 Fig. 9. (A) Phospholipase

Fig. 9. (A) Phospholipase A 2 activity of nigroviriditoxin B ( B ), compared to crotoxin B from Crotalus durissus terri cus ( C ), on the synthetic substrate 4-nitro-3-octanoyl- benzoic acid. Each point represents mean ± SD of triplicates. The differences between both enzymes are statistically signi cant ( p < 0.05; Student's t -test) at 2.5, 5, and 10 mg. (B) Myotoxic activity of nigroviriditoxin B (Ngvtx B) and crotoxin B (Ctx B) after intramuscular injection of 20 mg in mice. Plasma creatine kinase activity was deter- mined 3 h after injection. Phosphate-buffered saline (PBS) was injected to a control group of mice. Each bar represents mean ± SD of ve mice. Differences between all groups are statistically signi cant ( p < 0.05; ANOVA, followed by TukeyeKramer test).

indicate that nigroviriditoxin B displays a similar pro le of activities as crotoxin B, although with a general lower potency. Finally, nigroviriditoxin B was strongly recognized by equine antibodies

B was strongly recognized by equine antibodies Fig. 10. Modulation of the phospholipase A 2 activity

Fig. 10. Modulation of the phospholipase A 2 activity of nigroviriditoxin B by nigro- viriditoxin A. Enzyme activity of the B subunit alone, or in the presence of A subunit at the indicated molar ratios, was assayed on 4-nitro-3-octanoyl-benzoic acid. Each bar represents mean ± SD of triplicate assays. The difference in enzymatic activity of mixtures at both ratios, compared to nigroviriditoxin B alone is statistically signi cant ( p < 0.05; ANOVA, followed by Tukey eKramer test).

( p < 0.05; ANOVA, followed by Tukey e Kramer test). Fig. 11. Immunochemical cross-recognition of

Fig. 11. Immunochemical cross-recognition of nigroviriditoxin B by an equine anti- venom to Crotalus d. terri cus , evaluated by ELISA. Nigroviriditoxin B (A) , nigrovir- iditoxin A (B) , or crotoxin B (C) were adsorbed onto solid-phase and the binding of antibodies from antivenom ( C ) or normal horse serum ( B ) was detected as described in Materials and methods . Each point represents mean ± SD of triplicate wells.

from an antivenom raised against the venom of C. d. terri cus , demonstrating its antigenic similarity with crotoxin B, in agree- ment with their high sequence identity. From a structural point of view, it is noteworthy that the amino acid sequence differences between nigroviriditoxin B and crotoxin B, consisting of 24 out of 122 positions, do not predict any major deviations in the modeled three-dimensional structure of the former. In the crystal structure of crotoxin analyzed by Faure et al. (2011) , 22 amino acid residues of the B subunit were found to be engaged in its interaction with the A subunit. Of these, 18 are conserved in nigroviriditoxin B. Given that nigroviriditoxin A was shown to form a complex with crotoxin B, the high conservation of amino acid residues between the latter and nigroviriditoxin B would also predict a high conservation of key residues in the A subunits from both snake species. Among the four amino acid residues relevant for heterodimerization that differ between cro- toxin B and nigroviriditoxin B, the Trp 61 /Ser substitution (Trp 70 /Ser in the numbering system used by Faure et al. (2011) ) deserves special attention. Trp 30 and Trp 61 are considered to be critical to the stability and toxicity of the crotoxin complex, by establishing intermolecular contacts with the b-chain of crotoxin A ( Faure et al., 2011 ). Trp 30 is conserved in nigroviriditoxin B, but Trp 61 is replaced by Ser, and this could in uence either its binding af nity to nigroviriditoxin A, its toxicity, or both. Interestingly, inspection of the multiple alignment of PLA 2 s homologous to nigroviriditoxin B ( Fig. 4 ), shows that Trp 61 ( ¼ Trp 70 ) is exclusively present in the enzymes from rattlesnake species, whereas nigroviriditoxin B and the two PLA 2 s from Old World species (positioned basal to the rattlesnake group in the phylogenetic tree; Fig. 5 ), present Ser at this position. Considering that nigroviriditoxin B showed weaker toxicity than crotoxin B, it is therefore tempting to speculate that the mutation leading to the emergence of Trp 61 ( ¼ Trp 70 ) instead of

B. Lomonte et al. / Toxicon 93 (2015) 144 e154

153

B. Lomonte et al. / Toxicon 93 (2015) 144 e 154 153 Fig. 12. (A) Modeled

Fig. 12. (A) Modeled three-dimensional structure of nigroviriditoxin B (red-orange), superposed to the crystal structure of crotoxin B (green; PDB access code 3R0L; Faure et al., 2011 ). (B) The superposed structures are shown with the side chains of the 24 amino acid residues that differ, represented as sticks with their parent colors. These differences are highlighted with the same colors in the alignment shown. The twenty-two amino acid residues proposed to participate in the interaction of crotoxin B with its A-subunit ( Faure et al., 2011) are pointed out in the alignment by triangles. Positions indicated by black triangles are conserved in nigroviriditoxin B, whereas pink triangles indicate four amino acid positions that differ. Electrostatic surface potentials of the nigroviriditoxin B model (C) and crotoxin B (D) are represented with blue (positive) and red (negative) colors. Molecular representations were prepared with DS ViewerPro (Accelrys) software. (For interpretation of the references to color in this gure legend, the reader is referred to the web version of this article.)

Ser in a common ancestor to Crotalus and Sistrurus rattlesnakes might be related to the enhancement of toxic potency. Although toxic PLA 2 s having homology with crotoxin B chain have been found in a variety of viperid snake venoms ( Chen et al.,

2004 ), the simultaneous presence of A and B subunits of crotoxin or Mojave toxin has been described so far only in rattlesnakes, i.e. genera Crotalus and Sistrurus , among New World pitvipers. Our description of a crotoxin-like heterodimer in the venom of

B. nigroviridis represents, to the best of our knowledge, the rst

nding of such a complex in a non-rattlesnake New World pitviper

venom. This raises a number of intriguing issues regarding the molecular evolution of this neurotoxic trait. As evidenced by its

lethal potency, the toxicity of nigroviriditoxin is considerably lower than that of crotoxin and Mojave toxin, suggesting that it may represent a less differentiated, and hence less toxic stage in the molecular evolution towards neurotoxicity, although the structural paradigm of this heterodimeric design is clearly already present in

B. nigroviridis venom. These ndings suggest that the emergence of

this particular structural trait, involving a non-covalent A þ B complex between an acidic, non-toxic chaperone component and a toxic, basic PLA 2 , occurred earlier than previously thought in the evolution of pit viper venoms.

Con ict of interests statement

The authors declare that there are no con icts of interest related to this study.

Acknowledgments

We thank the International Centre for Genetic Engineering and Biotechnology (ICGEB, Italy; CRP/COS13-01), Vicerrectoría de Investigaci on, Universidad de Costa Rica (UCR; 741-B4-100 and 741-B3-760), CYTED (BioTox, P211RT0412), and Ministerio de Ciencia e Innovaci on (currently, Ministerio de Economía y Com- petitividad, Madrid; BFU2010-17373) for their nancial support, as well as Dr Mahmood Sasa for fruitful discussions on this study, and Dr Andreimar Soares for providing crotoxin B. The valuable contribution of Jazmín Arias, Danilo Chac on and Aar on G omez at the serpentarium of Instituto Clodomiro Picado is also gratefully acknowledged.

Transparency document

Transparency document related to this article can be found online at http://dx.doi.org/10.1016/j.toxicon.2014.11.235 .

154

B. Lomonte et al. / Toxicon 93 (2015) 144 e154

References

matics 22, 195 e 201. Biasini, M., Bienert, S., Waterhouse, A., Arnold, K., Studer, G., Schmidt, T., Kiefer, F., Gallo, T., Bertoni, M., Bordoli, L., Schwede, T., 2014. SWISS-MODEL: modelling protein tertiary and quaternary structure using evolutionary information. Nucleic Acids Res. http://dx.doi.org/10.1093/nar/gku340 . Calvete, J.J., Ju arez, P., Sanz, L., 2007. Snake venomics. Strategies and applications.

^

nez, V., Arce, V., Guti errez, J.M., Lomonte, B., 2004. Structural and functional characterization of myotoxin I, a Lys49 phospholipase A 2 homologue from the venom of the snake Bothrops atrox. Toxicon 44, 91 e101. Sampaio, S.C., Hyslop, S., Fontes, M.R.M., Prado-Franceschi, J., Zambelli, V.O., Magro, A., Brigatte, P., Gutierrez, V.P., Cury, Y., 2010. Crotoxin: novel activities for a classic beta-neurotoxin. Toxicon 55, 1045 e1060 . Savage, J.M., 2002. The Amphibians and Reptiles of Costa Rica. The University of Chicago Press, Chicago. Soares, A.M., Guerra-S a, R., Borja-Oliveira, C., Rodrigues, V.M., Rodrigues-Simioni, L., Rodrigues, V., Fontes, M.R.M., Lomonte, B., Guti errez, J.M., Giglio, J.R., 2000. Structural and functional characterization of BnSP-7, a lysine-49 myotoxic phospholipase A 2 homologue from Bothrops neuwiedi pauloensis venom. Archs. Biochem. Biophys. 378, 201 e209 . Sol orzano, A., 2004. Serpientes de Costa Rica. Editorial INBio, San Jos e . Tamura, K., Stecher, G., Peterson, D., Filipski, A., Kumar, S., 2013. MEGA6: molecular evolutionary genetics analysis version 6.0. Mol. Biol. Evol. 30, 2725 e 2729 . Van Devender, T.R., Conant, R., 1990. Pleistocene forest and copperheads in the Eastern United States, and the historical biogeography of New World Agkis- trodon . In: Gloyd, H.K., Conant, R. (Eds.), Snakes of the Agkistrodon Complex: a Monographic Review. Society for the Study of Amphibians and Reptiles, St. Louis, pp. 601 e614 . Wüster, W., Peppin, L., Pook, C.E., Walker, D.E., 2008. A nesting of vipers: phylogeny and historical biogeography of the Viperidae (Squamata: Serpentes). Mol. Phylogenet. Evol. 49, 445 e 459 .

~