BnP1, a novel P-I metalloproteinase from Bothrops neuwiedi venom: Biological effects benchmarking relatively to jararhagin, a P-III SVMP

ARTICLE IN PRESS

Toxicon 51 (2008) 54–65 www.elsevier.com/locate/toxicon

BnP1, a novel P-I metalloproteinase from Bothrops neuwiedi venom: Biological effects benchmarking relatively to jararhagin, a P-III SVMP$ C. Baldoa, I. Tanjonia, I.R. Leo´nb, I.F.C. Batistac, M.S. Della-Casaa, P.B. Clissaa, R. Weinlichd, M. Lopes-Ferreirae, I. Lebrunc, G.P. Amarante-Mendesd, V.M. Rodriguesf, J. Peralesb, R.H. Valenteb, A.M. Moura-da-Silvaa, a

Laborato´rios de Imunopatologia, Instituto Butantan, Av. Vital Brasil, 1500—05503-900, Sa˜o Paulo, SP, Brasil b Laborato´rio de Toxinologia, FIOCRUZ, Rio de Janeiro, Brasil c Bioquı´mica e Biofı´sica, Instituto Butantan, Sa˜o Paulo, Brasil d Laborato´rio de Biologia Celular e Molecular, ICB/USP, Sa˜o Paulo, Brasil e Toxinologia Aplicada, Instituto Butantan, Sa˜o Paulo, Brasil f Laborato´rio de Quı´mica de Proteı´nas, UFU, Minas Gerais, Brasil Received 28 June 2007; received in revised form 9 August 2007; accepted 10 August 2007 Available online 14 August 2007

Abstract Snake venom metalloproteinases (SVMPs) have been extensively studied and their effects associated with the local bleeding observed in human accidents by viper snakes. Representatives of P-I and P-III classes of SVMPs similarly hydrolyze extracellular matrix proteins or coagulation factors while only P-III SVMPs induce significant hemorrhage in experimental models. In this work, the effects of P-I and P-III SVMPs on plasma proteins and cultures of muscle and endothelial cells were compared in order to enlighten the mechanisms involved in venom-induced hemorrhage. To reach this comparison, BnP1 was isolated from B. neuwiedi venom and used as a weakly hemorrhagic P-I SVMPs and jararhagin was used as a model of potently hemorrhagic P-III SVMP. BnP1 was isolated by size exclusion and anion-exchange chromatographies, showing apparent molecular mass of approximately 24 kDa and sequence similarity with other members of SVMPs, which allowed its classification as a group P-I SVMP. The comparison of local effects induced by SVMPs showed that BnP1 was devoid of significant myotoxic and hemorrhagic activities and jararhagin presented only hemorrhagic activity. BnP1 and jararhagin were able to hydrolyze fibrinogen and fibrin, although the latter displayed higher activity in both systems. Using HUVEC primary cultures, we observed that BnP1 induced cell detachment and a decrease in the number of viable endothelial cells in levels comparable to those observed by treatment with jararhagin. Moreover, both BnP1 and jararhagin induced apoptosis in HUVECs while only a small increase in LDH supernatant levels was observed after treatment with jararhagin, suggesting that the major mechanism involved in endothelial cell death is apoptosis. Jararhagin and BnP1 induced little effects on C2C12 muscle cell cultures, characterized by a partial detachment

$

Ethical statement: The paper represents a series of experiments carries out under the standard procedures of scientific ethics, including the care of experimental animals. All authors have read the manuscript and agree to its publication in Toxicon. Corresponding author. E-mail address: [email protected] (A.M. Moura-da-Silva). 0041-0101/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.toxicon.2007.08.005

ARTICLE IN PRESS C. Baldo et al. / Toxicon 51 (2008) 54–65

55

24 h after treatment and a mild necrotic effect as evidenced by a small increase in the supernatants LDH levels. Taken together, our data show that P-I and P-III SVMPs presented comparable effects except for the hemorrhagic activity, suggesting that hydrolysis of coagulation factors or damage to endothelial cells are not sufficient for induction of local bleeding. r 2007 Elsevier Ltd. All rights reserved. Keywords: Snake venom; Metalloproteinases; Hemorrhage; Endothelial cells; Myotoxicity

1. Introduction Snake venom metalloproteinases (SVMPs) have been extensively studied and considered as playing key roles in the development of venom-induced pathogenesis. In general, the action of SVMPs is related to proteolysis of extracellular matrix components (type IV collagen, laminin, fibronectin and nidogen), plasma proteins (fibrinogen, fibrin, Von Willenbrand factor, pro-thrombin) and cell surface proteins (integrins and cadherins). Moreover, SVMPs are also able to interact with receptors of platelets, endothelial cells and fibroblasts, activating or inhibiting the cellular response. These effects promote several pathologic alterations observed after envenoming such as hemorrhage, inhibition of platelet aggregation, coagulopathy, myonecrosis and inflammatory response (reviewed by Mourada-Silva et al., 2007; Gutie´rrez et al., 2005; White, 2005). The functional diversity exhibited by SVMPs is directly related to their structural complexity. These toxins belong to the metzincin super-family of zincdependent peptidases, a diverse and expansive group present on both prokaryotes and eukaryotes, which encloses also metalloproteinases important for mammalian functions as matrix-degrading metalloproteinases (MMPs) and ADAMs (A Disintegrin And Metalloproteinase) (Bode et al., 1993; Stocker and Bode, 1995). SVMPs are synthesized as multidomain zymogen proteins and are grouped in classes and subclasses from P-I to P-IV according to their domain organization (reviewed by Fox and Serrano, 2005). P-I and P-III SVMPs are the most abundant groups in viper venoms, which conserve the catalytic domain in the mature form. P-I SVMPs are composed in the mature form only by the catalytic domain and are generally fibrino(geno)lytic enzymes, weakly hemorrhagic. In opposition, P-III SVMPs are frequently potent hemorrhagins and their structures differ from P-I SVMPs by the addition of disintegrin-like and cysteine-rich domains carboxy-terminally to the catalytic domain.

Neuwiedase, isolated from Bothrops neuwiedi venom (Rodrigues et al., 2000), and jararhagin, isolated from Bothrops jararaca venom (Paine et al., 1992), respectively, represent P-I and P-III SVMPs. Neuwiedase, a 22 kDa metalloproteinase, is able to degrade fibrinogen, fibrin, type I collagen, fibronectin, laminin and induces inflammatory reaction. Neuwiedase is devoid of hemorrhagic activity on skin tests, but induced bleeding when applied onto the mouse cremaster muscle in intravital microscopy experiments. Besides the hemostatic effects, neuwiedase also induced a myotoxic reaction in mice model (Rodrigues et al., 2001). As representative of P-III SVMPs, jararhagin is a potent hemorrhagin able to inhibit collagen-induced platelet aggregation (Kamiguti et al., 1996a). Jararhagin hydrolyzes fibrinogen (Kamiguti et al., 1994) and Von Willebrand factor (Kamiguti et al., 1996b), induces a strong pro-inflammatory response increasing the number of rolling leukocytes within the capillary vessels (Clissa et al., 2006), inducing accumulation of polymorphonuclear cells at the site of injection (Costa et al., 2002) and increasing local and systemic cytokine levels (Clissa et al., 2001; Laing et al., 2003). Jararhagin also affects endothelial cell growth by promoting apoptosis (Tanjoni et al., 2005) or releasing angiostatin-like peptides (Ho et al., 2002). Attempting to understand the differences between biological effects of P-I and P-III SVMPs, the aims of this study were to isolate neuwiedase from B. neuwiedi venom and to compare its biological effects with jararhagin activity in parallel experiments. Two analogues of neuwiedase were isolated from the venom of B. neuwiedi (BnP1 and BnP2) and the biological effects of one of them, BnP1, towards hemostatic system and cultures of muscle and endothelial cells, were comparable to those of jararhagin, except for the hemorrhagic activity, suggesting that hydrolysis of coagulation factors or damage of endothelial cells are not sufficient for induction of local bleeding.

ARTICLE IN PRESS 56

C. Baldo et al. / Toxicon 51 (2008) 54–65

2. Material and methods 2.1. Isolation of toxins Venoms of B. jararaca and B. neuwiedi were collected from snakes kept under captivity at Instituto Butantan herpentarium (Laborato´rio de Herpetologia). B. neuwiedi venom was fractionated by size exclusion and anion-exchange chromatographies using the FPLC system. Fractions containing metalloproteinases were screened according to their fibrinolytic activity as described below. Crude venom (13 mg) was resuspended into 20 mM Tris–HCl pH 8.0 buffer and applied to a Superdex 75 HR 10/30 column, the fractions were eluted with the same buffer at a flow rate of 0.5 mL/min. The peak containing fibrinolytic activity was further chromatographed on a Mono-Q column (HR 5/5) using 20 mM Tris/HCl buffer, pH 8.0, containing 1 mM CaCl2. Elution was performed by a linear NaCl gradient (0–1 M in 30 mL) at a flow rate of 1.0 mL/min in the same buffer. All steps were carried out at room temperature and monitored at 280 nm. Jararhagin was isolated from B. jararaca snake venom as previously described (Paine et al., 1992; Moura-da-Silva et al., 2003). Purity of toxins was evaluated by SDS-PAGE 12.5%, under reducing conditions (Laemmli, 1970). 2.2. Protein sequencing The N-terminal amino acid sequence of the BnP1 was determined by automated Edman degradation (Edman and Begg, 1967) using a Shimadzu Protein Sequencer (PPSQ-21). The protein bands from SDS-PAGE were submitted to in gel digestion with trypsin (Leon et al., 2007) and eluted peptides sequenced by mass spectrometry. For this purpose, 1 mL of the eluted peptides were applied to a ProteCol C18 column (SGE, Australia) at 4.5 mL/min of 0.1% (v/v) formic acid in water (A) or in acetonitrile (B), using the following gradient: 5% B for 5 min, 40% B at 35 min, 60% B at 45 min, 80% B at 48 min. Eluting peptides in the column effluent were directly electrosprayed into a LCQ Deca XP Plus ion trap spectrometer (Thermo Finnigan, USA) for analysis. Spectra of eluting peptides were acquired in a data-dependent fashion by first acquiring a full MS scan from m/z 400 to 2000 followed by a ZoomScan mode between m/z 400 and 2000 of the most intense ions. MS/MS scans were acquired using an activation qz of 0.25,

for 30 ms and 35% normalized collision energy. Alternatively, 0.3 mL of peptide sample and 0.3 mL of matrix solution [saturated a-cyano-4-hydroxycinnamic acid in 50% acetonitrile in 0.1% (v/v) trifluoroacetic acid in water] were co-crystallyzed on the sample plate and MALDI-MS was performed on a 4700 Proteomics Analyzer with version 3.0 software (Applied Biosystems, USA). MS spectra were acquired in positive ion reflector mode with 1250 laser shots per spot, processed with default calibration and the six most intense ions submitted to fragmentation. MS/MS spectra were acquired with 3000 laser shots and 1 keV collision energy with CID off (1e8 Torr). Acquired data were manually interpreted and obtained sequences submitted to the NCBI non-redundant database using the Blastp software (Altschul et al., 1997). 2.3. Hemorrhagic and myotoxic activities Hemorrhagic activity was assessed using samples of 2.5 and 10 mg of jararhagin or 10 and 50 mg BnP1 (diluted in 50 mL sterile PBS), injected i.d. into the dorsal skin of Swiss mice (n ¼ 5). Control mice received only sterile PBS. After 3 h, the animals were sacrificed, the dorsal skin removed and the extent of hemorrhagic spots analyzed. The hemorrhagic activity was also determined using intravital microscopy by transillumination of mice cremaster muscle after topical application of 1 mg of jararhagin or 50 mg of BnP1 diluted in 20 mL sterile PBS. Microcirculation alterations were observed during 30 min by light microscopy (Axioskope, Carl-Zeiss). The myotoxic activity was assayed using 50 mg of jararhagin and BnP1 or 10 mg of B. neuwiedi venom, injected i.m. in Swiss mice (n ¼ 5). After 3 h, the animals were bled and the sera were assayed for creatine-kinase activity with a commercial kit (Bioclin). Mean values were compared using the Student’s t-test. The conducts and procedures involving animal experiments were approved by the Instituto Butantan Committee for Ethics in Animal Experiments (License number CEUAIB 191/2004). 2.4. Fibrino(geno)lytic activity Fibrinolytic activity was assayed by fibrin-plate method, as previously described (Jespersen and Astrup, 1983). Briefly, a fibrin-agarose gel was prepared by mixing 1 mg/mL human fibrinogen

ARTICLE IN PRESS C. Baldo et al. / Toxicon 51 (2008) 54–65

(Calbiochem) with a pre-heated solution of 2% agarose in 50 mM Tris–HCl pH 7.3 buffer containing 200 mM NaCl, 50 mM CaCl2 and 2 U/mL thrombin. Samples (0.48–20 mg/mL) were applied to wells of the solidified gel and incubated at 37 1C for 18 h, and then the hydrolyzed area was measured. Fibrinogenolytic activity was carried out by incubating bovine fibrinogen solution with the metalloproteinases (1:20 toxin–substrate weight rate) in 20 mM Tris–HCl, pH 8.0 buffer containing 150 mM NaCl and 1 mM CaCl2, for increasing periods of time, at 37 1C. The reaction products were analyzed by 12.5% SDS-PAGE under reducing conditions. 2.5. Cell culture assays Endothelial cells were obtained from human umbilical cord veins (HUVECs) by type IV collagenase (Worthington) digestion as described by Jaffe et al. (1973) and cultured on gelatin-coated plastic dishes in RPMI 1640 medium supplemented with 10% (v/v) FBS, 2 mM L-glutamine, 1 mM sodium pyruvate, 100 UI/mL penicillin, 100 mg/mL streptomycin, 50 mM 2-mercaptoethanol, 5 U/mL heparin and 20 ng/mL bFGF. The HUVEC cells were seeded into 24-well microplates, at an approximate initial density of 0.5  105 cells/well, in the same medium and used for all experiments after 24 h culture. C2C12 myoblasts were cultivated in RPMI 1640 medium supplemented with 10% (v/v) FBS, 2 mM L-glutamine, 2 mM sodium pyruvate, 1 mM non-essential amino acids, 100 UI/mL penicillin and 100 mg/mL streptomycin. The cells were seeded into 24-well microplates, at an approximate initial density of 2  103 cells/well, in the same medium. After reaching near confluence, usually in 1–2 days, growth media were replaced by differentiation media, which consisted of the same medium supplement with 1% FBS. After six additional days of culture, when a vast proportion of long multinucleated myotubes was observed, cells were used in all assays. Cell viability and cell adhesion of cultures treated with the metalloproteinases were evaluated by MTT assay as previously described (Tanjoni et al., 2005). Cell necrosis was evaluated by lactic dehydrogenase (LDH) release in culture supernatants after centrifuging at 240 g during 5 min, using a colorimetric end-point assay (LDH-FS DiaSys-Diagnostic Systems International) according to manufacturer instructions.

57

Apoptosis was accessed by the DNA content of cells stained with Hypotonic Fluorescent Solution (50 mg/mL P-I, 0.1% Triton X-100 and 0.1% sodium citrate), analyzed by FACScalibur (Becton-Dickinson, San Jose, CA, USA). DNA fragmentation was quantified by cell cycle analysis of total DNA content as previously described (Nicolleti et al., 1991). Using the FSC  SSC parameters, debris were gated out and 10,000 events were counted. Acquisition and analysis were performed with Cell Quest software (BectonDickinson) and the results represent the percentage of the population gated in the region corresponded to sub-G1 phase (fractional DNA content). Samples were run in triplicates and the results expressed as mean7sd (Tanjoni et al., 2005). For all experiments, mean values were compared using the Student’s t-test. 3. Results 3.1. Isolation of BnP1 and BnP2 and partial sequences Size exclusion chromatography resolved B. neuwiedi venom in six major fractions (Fig. 1A) and the fibrinolytic activity was mainly concentrated in the fraction III. When submitted to anion-exchange chromatography at pH 8.0, fraction III was resolved into three peaks, two of them with fibrinolytic activity: one peak, IIIa/IIIb, did not bind to the column and the other, IIIc, was bound and eluted with 0.15 M NaCl, suggesting their different isoelectric properties (Fig. 1B). These proteins presented similar apparent molecular mass, determined by SDS-PAGE, of approximately 24 kDa under reducing conditions (Fig. 1C). Both proteins also showed a single band in SDS-PAGE run under nonreducing conditions (data not shown), indicating that they exist as single-chained proteins. The homogeneity of BnP2 was only approximately 90% and therefore, this protein was used only for structural characterization, being the biological data relative only to BnP1. The partial amino acid sequence of the isolated proteins was carried out by mass spectrometry and by N-terminal determination reached by Edman degradation. Using these two approaches, approximately 40% of sequence from fraction IIIa and 34% of fraction IIIc were covered. Sequences were then aligned and compared with the original sequence of neuwiedase (Rodrigues et al., 2000). As shown in

ARTICLE IN PRESS

IIId

IIIa - BnP1 IIIb

1

0.4

0.8 0.6 0.4

0.75

0.3 0.5 0.2 A280nm 0.25 NaCl (%)

0.1

0.2

0 31

BnP2

BnP1

21 Time (min)

IIId

11

IIIc

IIIb

IIIa

VI

0 1

51

V

41

IV

III

31 Time (min)

II

21

I

11

Bn V.

1

MW

0.0

NaCl (M)

V

VI

IV

0.5

A280 (nm)

A280 (nm)

1.0

III

I II

1.2

IIIc - BmP2

C. Baldo et al. / Toxicon 51 (2008) 54–65

58

116 -66.2 -45.0 -35.0 -25.0 -18.4 -Venom

Size exclusion

Anion exchange

Isolated toxins

Fig. 1. Isolation of BnP1 and BnP2. Samples of 13 mg of Bothrops neuwiedi venom were applied on Superdex 75 HR 10/30 column, equilibrated with 20 mM Tris–HCl and 1 mM CaCl2 pH 8.0. Fractions of 1.0 mL were collected at a flow rate of 0.5 mL/min, monitored at A280 (nm) (A). Fraction III of size exclusion chromatography was applied to a Mono-Q HR 5/5, equilibrated with 20 mM Tris–HCl pH 8.0 plus 1 mM CaCl2. Bound proteins were eluted with 30 mL of a 0–1 M NaCl linear gradient and fractions of 1.0 mL were collected at a flow rate of 1.0 mL/min, monitored at A280 (nm) (B). Whole B. neuwiedi venom, fractions of size exclusion chromatography (I–VI), fractions of the ion exchange column (IIIa–IIId) and isolated toxins (BnP1 and BnP2) were analyzed by SDS-PAGE under reducing conditions and apparent molecular mass calculated according to migration of molecular mass markers (MW) (C).

Fig. 2, the two partial sequences revealed a high percentage identity showing only one replacement between Asn and Asp at position 30. However, when the toxins were compared with neuwiedase (gi 6760464), a greater divergence was observed, with identity of only 69% under the covered region of both sequences. These data allowed us to conclude that the isolated proteins are not isoforms of the originally isolated neuwiedase but consist in new SVMP members that we designated as BnP1 and BnP2. The next step was then to run a Blastp search (Altschul et al., 1997) for similar proteins to BnP1 and BnP2. The program predicted the presence of putative conserved domains of reprolysin and indicated the highest similarity with insularinase (gi 52426579, 91% identity), followed by type II and type I metalloproteinases from Bothrops asper venoms (gi 82466485, 82% identity

and gi 82466487, 82% identity, respectively), BaP1 (gi 46395679, 81% identity) and neuwiedase (gi 6760464, 69% identity). Under the covered region, the identity with jararhagin, a P-III SVMP (gi 62468), was only 50% (Fig. 2). Although the catalytic motif was not covered by the sequences resolved in the new toxins, the molecular size, presence of fibrinolytic activity and high percentual of identity with SVMPs indicate that they are new members of P-I class of SVMPs present in B. neuwiedi venom. Despite of similarities between BnP1 and BnP2, they might have different isoelectric properties, as deduced from their behavior on ion exchange elution. Total amino acid composition of isolated proteins showed that BnP1 exhibits a higher concentration of Lys, Arg and Ser whereas BnP2 shows a higher concentration of Met (data not

ARTICLE IN PRESS C. Baldo et al. / Toxicon 51 (2008) 54–65

10 BnP1 BnP2 Insularinase Type II (Ba) Type I (Ba) BaP1 Neuwiedase Jararhagin

40

50

60

70

80

100

110

120

130

140

150

160

FGEWRERDJJPR-------------------------------------------------------------------............-------------------------------------------------------------------........LL..ISHDHAQLLTTIVFDQQTIGLAYTAGMCDPRQSVAVVMDHSKKNLRVAVTMAHELGHNLGMDHD-DTC ........LL..ISHDHAQLLTTIVFDNYVIGITEFGKMCDPKLSVGVVRDHSEINLQVAVAMAHELGHNLGMYHDGNQC ........LL..ISHDHAQLLTAVVFDGNTIGRAYTGGMCDPRHSVGVVRDHSKNNLWVAVTMAHELGHNLGIHHDTGSC ........LL..ISHDHAQLLTAVVFDGNTIGRAYTGGMCDPRHSVGVVRDHSKNNLWVAVTMAHELGHNLGIXHDTGSC ........LLR.KSHDNAQLLTAIDFNGNTIGRAYLGSMCNPKRSVGIVQDHSPINLLVGVTMAHELGHNLGMEHDGKDC .A...KT.LLT.KKHDNAQLLTAIDFNGPTIGYAYIGSMCHPKRSVGIVQDYSPINLVVAVIMAHEMGHNLGIHHDTGSC 170

BnP1 BnP2 Insularinase Type II (Ba) Type I (Ba) BaP1 Neuwiedase Jararhagin

30

SQIKFKP-SYIELAVVADHGMFTKYNSNJNTJR-JVHEMVNTVDGFFR----------------------------TJTS ---------....................D...–-.............SMNVDASJANJEVWSK------------.... E.Q..S.-R...................L..I.TR.......LN.....V.....L..L.....KDLIKVEKDSSK.L.. E.QR.S.-R...........I.......L..I.TR....L...N..Y..V..T..L.SL.....KDLIKVEKDSSK.L.. E.QR.S.-R...........I.......L.TI.TR....L...N..Y..VD.H.PL..L.....QDLIKVQKDSSK.LK. -.QR.S.-R...........I.......L.TI.TR....L...N..Y..VD.H.PL..L.....QDLIKVQKDSSK.LK. Q.RF.PQ-R....VI...RR.Y.....DS.KI.TR...L....N...........L..L.....KDLIKVEKDSSK.L.. E.QRYD.YK...FF..V.Q.TV..N.GDL.KIKARMY.LA.I.NEI..YLYMHVALVGL.I..NGDKITVKPDVDY.LN. 90

BnP1 BnP2 Insularinase Type II (Ba) Type I (Ba) BaP1 Neuwiedase Jararhagin

20

59

180

190

200

-----SCJMASTJSK--------------------HNPQCJJNQPJ ---------------------------------------------TCGAK..I....I..GLSFEFSKCSQNQYQTYLTD.....IL.K.. HCDAA..I..D.LREVLSYEFSDCSQNQYETYLTK.....IL.E.. SCGAK..I...VL..VLSYEFSDCSQNQYETYLTN.....IL.K.. SCGAK..I...VL..VLSYEFSDCSQNQYETYLTN.....IL.K.LCGASL.I.SPGLTDGPSYEFSDCSKDYYQTFLTN....------SCGDYP.I.GP.I.NEPSKFFSNCSYIQCWDFIMN...E.II.E..

91% 82% 82% 81% 69% 50%

identity identity identity identity identity identity

Fig. 2. Sequence alignment of BnP1 and BnP2 with SVMPs. Partial BnP1 and BnP2 sequences were obtained by mass spectrometry (J: Ile or Leu) and Edman degradation (underlined residues) and aligned against other SVMPs using CLUSTAL W multiple alignment package (Combet et al., 2000). Dashes represent gaps to improve the sequence alignment. Gray box corresponds to regions covered by the partial BnP1/BnP2 sequences where identical residues are indicated by dots and the SVMPs zinc-binding motif is shown in opened rectangle.

shown). However, the differences between the amino acid composition of BnP1 and BnP2 were not detected on the fragments already sequenced, they probably will be confirmed after complete sequences are achieved. 3.2. Comparison between BnP1 and jararhagin hemorrhagic activity and hydrolysis of clotting factors The next step was to compare the major activities of BnP1 with jararhagin effects on hemostatic components. Hemorrhagic activities of BnP1 and jararhagin were evaluated in two different mice models: by i.d. injection of toxins on the dorsum of mice followed by visual inspection and dropping different doses in mice cremaster muscle analyzed by intravital microscopy. Injection of 2.5 mg jararhagin in the mice dorsal skin induced a strong hemorrhagic spot (Fig. 3A) that increased even further in doses of 10 mg (Fig. 3B). In opposition, the doses of 10 mg BnP1 did not induce any

detectable hemorrhage (Fig. 3E), which was detected as a weak spot only with doses of 50 mg of BnP1 (Fig. 3E). Similar results were observed by intravital microscopy: 50 mg BnP1 did not induce any detectable change in the capillary vessels or in the surrounding muscle cells (Fig. 3H). In the same model, 1 mg jararhagin was enough to induce detectable damage in the capillary network with blood extravasation (Fig. 3D, dashed lane) and to activate endothelium of post-capillary venules with increase in the number of rolling leukocytes (Fig. 3D, arrow). The fibrinolytic activity assayed by fibrin plates showed that BnP1 and jararhagin were able to degrade fibrin in a dose-dependent manner. However, considering the molar basis, jararhagin showed a fibrinolytic activity approximately 5 times higher than BnP1 (Fig. 4A). The proteolytic activities of BnP1 and jararhagin were also investigated by hydrolysis of human fibrinogen followed by fragment analysis in SDS-PAGE (Fig. 4B). When the reaction products were analyzed, it was

ARTICLE IN PRESS C. Baldo et al. / Toxicon 51 (2008) 54–65

60

JARARHAGIN

A

B

2.5 µg

C

10 µg

D

1 µg-10 min

Before toxin adm.

BnP1

F

E

10 µg

G

50 µg

H

Before toxin adm.

50 µg - 10 min

Fig. 3. Hemorrhagic activity of BnP1 and jararhagin. Samples of 2.5 mg (A) or 10 mg (B) of jararhagin, and 10 mg (E) or 50 mg (F) of BnP1 were injected (i.d.) on the dorsum of mice followed by visual inspection of hemorrhagic lesion. Hemorrhagic activity was also analyzed by intravital microscopy of mouse cremaster muscle 20 min after treatment with 1 mg jararhagin (D) or 50 mg BnP1 (H) compared to the same optical site before application of jararhagin or BnP1 (C and G, respectively). Photographs were obtained from digitized images on the computer monitor. The dashed lane represents the hemorrhagic area and the arrow indicates the rolling leukocytes induced by jararhagin. The data are representative of observations in three mice analyzed on different days.

evident that BnP1 was able to completely degrade the fibrinogen a chain within 1 h. Nevertheless, the proteolysis of a chain by jararhagin already occurred at 15 min. The hydrolysis of b chain by jararhagin also occurred in earlier times than by BnP1. The g chain proteolysis was not observed by any toxin. The fibrinogenolytic activity of both toxins was completely abolished when they were pre-incubated with 20 mM EDTA (Fig. 4B, lanes 7). 3.3. Action of SVMPs on HUVECs and C2C12 cells It has already been reported that jararhagin induces detachment and loss of viability of endothelial cells (Tend cell line) by apoptosis (Tanjoni et al., 2005) and that neuwiedase is able to induce myonecrosis in mice (Rodrigues et al., 2001). In this work we compared the effects of BnP1 and jararhagin on primary cultures of human umbilical vein endothelial cells (HUVECs) and C2C12 myotube cultures. Very similar effects of jararhagin and BnP1 were observed on cell cultures: both jararhagin and BnP1 affected the cell adhesion of HUVEC cultures in time-dependent manner. Effects on HUVEC cell adhesion could already be observed 6 h after treatment. After 24 h, a drastic decrease in the number of adherent cells was observed and

almost none of the cells remain attached after 48 h (Fig. 5A). The viability of HUVEC cell cultures detected by MTT assays after treatment with jararhagin and BnP1 showed a slight decrease after 6 h treatment, after 24 h incubation the number of viable cells was reduced and almost no viable cells were observed after 48 h of treatment (Fig. 5B). On the opposite way, only partial detachment of C2C12 cells was observed after treatment with BnP1 or jararhagin (Fig. 5C) and no significant effect on C2C12 cultures viability was detected by treatment of both SVMPs (Fig. 5D). The mechanisms involved in reduction of cell viability were also compared showing that treatment of HUVEC but not C2C12 cultures with jararhagin and BnP1 resulted in apoptosis. Apoptosis of HUVECs was not observed after 6 h treatment either with BnP1 or jararhagin and only a discrete increase was observed after 24 h (12.5% apoptosis by jararhagin and 11.4% by BnP1). Consistent values of apoptosis were observed only 48 h after treatment with jararhagin (50%) and BnP1 (76%), as shown in Fig. 6A. C2C12 cultures submitted to the same treatments did not undergo apoptosis (Fig. 6C). On the other hand, only a mild increase in LDH concentration was observed in supernatants of C2C12 cultures treated with both toxins, in periods

ARTICLE IN PRESS C. Baldo et al. / Toxicon 51 (2008) 54–65

61

4. Discussion Fibrinolytic activity (mm2)

35

Jararhagin BnP1

30 25 20 15 10 5 0 1

10 Toxins (μg/mL) BnP1

JARARHAGIN

α β γ

1

2

3 4

5

6

7

1

2

3

4

5

6

7

Fig. 4. Fibrino(gen)olytic activity of BnP1 and jararhagin. (A) Fibrinolytic activity was calculated by measuring the hydrolysis halo induced by the toxins in fibrin-agarose plates. The experiments were carried out in duplicates and the data represent mean7sd of three independent experiments. (B) Samples of 20 mg human fibrinogen (1) were incubated with 1 mg jararhagin or BnP1 for different periods (2: 15 min; 3: 30 min; 4: 1 h; 5: 3 h; 6: 6 h) and hydrolysis analyzed by SDS-PAGE 12.5% gels under reducing conditions. Hydrolysis by toxins pre-incubated for 30 min with 20 mM EDTA is shown in lane (7). Gels were stained with Coomassie brilliant blue.

over 48 h (Fig. 6D) and, at the same periods, of HUVEC cultures treated with jararhagin (Fig. 6B), suggesting that necrosis, if involved in cell damage by SVMPs, occurs in minor proportions than apoptosis of HUVECs. The relevance of SVMPs myotoxicity was also assayed in vivo by detection of blood CK increase in mice injected i.m. with both toxins. As shown in Fig. 7, injection of mice gastrocnemius muscle with 50 mg jararhagin or BnP1 did not induce significant increase of serum CK levels when compared to PBS injection. Significant myotoxicity (po0.05) was observed only after injection of 10 mg B. neuwiedi venom, suggesting that BnP1 and jararhagin SVMPs do not play any role in direct venom-induced myotoxicity.

In this report we describe the isolation and partial characterization of two P-I SVMPs from B. neuwiedi venom and compare the biological effects of one of them (BnP1) with biological activities of jararhagin, a well-known SVMP from P-III class. The isolated proteins were submitted to partial protein sequencing revealing identity with the catalytic domain of SVMPs. The highest similarity was achieved with insularinase A, a prothrombin activator from Bothrops insularis (Modesto et al., 2005), followed by Type II metalloproteinase from Bothrops asper venom (gi 82466485, unpublished) and the lowest was with the catalytic domain of class P-III SVMPs, here represented by jararhagin. It is interesting to note that both insularinase and Type II Bothrops asper metalloproteinases derive from N-II cDNA sequences, which code for precursors of class P-II SVMPs, which includes the RGD disintegrins. However, all the peptides detected by sequencing were located on the catalytic domain and the molecular mass estimated for BnP1 and BnP2 corresponds to a class P-I SVMP, suggesting that BnP1 and BnP2 may be synthesized from a N-II cDNA but are processed and presented on B. neuwiedi venom as a P-I SVMP. The identity shared between BnP1/BnP2 and neuwiedase, a SVMP P-I from B. neuwiedi (Rodrigues et al., 2000), was of only 69% suggesting that they are not homologous toxins but may be derived from distinct copies of genes coding for SVMPs. In this regard, the evolution of SVMP gene family apparently involves several gene duplications and accelerated evolution of each gene copies responsible for the large functional and structural diversity observed for snake metalloproteinases (Moura-da-Silva et al., 1996). This variability may account for differences in venom composition observed according to phylogeny, ontogeny or geographical variations (Chippaux et al., 1991), supporting the findings of this paper showing the pre-dominance of BnP1 and BnP2 in our pool of venom instead of neuwiedase. Despite structural differences, biological activities of BnP1 (this paper) and BnP2 (data not shown) were very similar to neuwiedase, that is, they are weakly hemorrhagic metalloproteinases with preferential activity towards the Aa chain of fibrinogen, also with fibrinolytic activity (Rodrigues et al., 2000). However, the myotoxic activity of BnP1 differed from the action of neuwiedase since the

ARTICLE IN PRESS C. Baldo et al. / Toxicon 51 (2008) 54–65

0.5

0.6

0.4

0.5

Viability - OD 570 nm

HUVEC Adhesion - OD 570 nm

62

0.3 0.2

# #

0.1

*

#*

0.0

0.4 *#

0.3

#

0.2 0.1

#*

*

0.0 0

12

24

36

48

0

12

24

0.7

0.7

0.6

0.6

*

0.5

#

0.5 0.4 #*

0.3

*

0.2 0.1

Viability - OD 570 nm

C2C12 Adhesion - OD 570 nm

PBS Jararhagin (#) BnP1 (*)

0.0

36

48

0.4 0.3 0.2 0.1 0.0

0

12 24 36 Time after treatment (hours)

48

0

12 24 36 Time after treatment (hours)

48

Fig. 5. Effects of BnP1 and jararhagin on HUVEC and C2C12 cell adhesion viability. HUVEC (A, B) and C2C12 (C, D) cultures were treated with 800 nM jararhagin or BnP1 over a period of 48 h. In different times, adhesion (A, C) and viability (B, D) were evaluated as described in experimental procedures. The data are expressed as mean7sd of one representative experimental performed in triplicate. Significance was estimated by Student’s T test and the symbols correspond to po0.05, when BnP1 (*) or jararhagin (#) data were compared with PBS at the same experimental conditions.

latter induces myotoxic effects in mice models, evidenced by histologically and by the increase in plasma creatine-kinase levels (Rodrigues et al., 2001), but these effects were observed neither in mice injected with the former nor in cultures of C2C12 muscle cells with it incubated. It is important to note that in this study we evaluated the direct myotoxicity of BnP1, observed by early release of CK in vivo or cell cytotoxicity in muscle cell cultures. A number of studies have documented that hemorrhagic metalloproteinases induce acute muscle damage (Gutie´rrez and Rucavado, 2000). However, such myotoxicity does not depend on the direct action on muscle cells, but instead on an indirect mechanism, possibly related to ischemia due to impairment of tissue perfusion as a consequence of microvasculature damage. The biological activities of BnP1 and jararhagin were then compared in relation to their effects on hemostasis, attempting to understand the differ-

ences in the hemorrhagic disturbances induced by P-I and P-III SVMPs. It has been reported that many SVMPs, mainly from P-III class, present hemorrhagic activity. However, several SVMPs generally of P-I class are devoid or present very weak hemorrhagic activity (reviewed by Mourada-Silva et al., 2007). Thus, comparative studies using hemorrhagic and non-hemorrhagic SVMPs are still necessary to clarify the hemorrhagic mechanism of P-III SVMPs. For this purpose, comparison of jararhagin and BnP1 appears to be a good model since jararhagin is strongly hemorrhagic and BnP1 presents a very weak hemorrhagic activity. As already described, jararhagin cleaves a 23 kDa fragment from the C-terminal part of the Aa chains of fibrinogen, while b and g chains remain unaffected (Kamiguti et al., 1994a). In this work, we observed that BnP1 as well as jararhagin have preferential activity toward the Aa chain of fibrinogen and could be considered A-a fibrinogenases.

ARTICLE IN PRESS C. Baldo et al. / Toxicon 51 (2008) 54–65

100

*

80

#

60

#

40 20

LDH (U/L)

HUVEC Apoptosis (%)

PBS Jararhagin (#) BnP1 (*)

120

100

#*

0

80 # 60 40 20

*

0

0 12

24

36

48

0

100

120

80

100 LDH (U/L)

C2C12 Apoptosis (%)

63

60 40 20

*#

0

12

24

36

48

80 60 *#

40 20 0

0

12 24 36 Time after treatment (hours)

48

0

12 24 36 Time after treatment (hours)

48

Fig. 6. The mechanisms by which BnP1 and jararhagin on HUVEC and C2C12 affect cell viability. HUVEC (A, B) and C2C12 (C, D) cultures were treated with 800 nM jararhagin or BnP1 over a period of 48 h. In different times, % apoptosis (A, C) was evaluated as described in experimental procedures and cytotoxicity estimated by the release of LDH to the culture media (B, D). The data are expressed as mean7sd of one representative experimental performed in triplicate. Significance was estimated by Student’s T test and the symbols correspond to po0.05, when BnP1 (*) or jararhagin (#) data were compared with PBS at the same experimental conditions.

However, identification as either an a- or b-fibrinogenases might not be considered as absolute since usually there is significant degradation of alternative fibrinogen chains with increasing times of incubation (Swenson and Markland, 2005). In this regard, we also observed the proteolysis of Bb chain of fibrinogen by jararhagin and also by BnP1. However, Aa chain is more rapidly hydrolyzed than Bb and for both chains, the hydrolysis provoked by jararhagin was faster than BnP1 hydrolysis rates. Fibrin was also digested by BnP1 and jararhagin, but jararhagin fibrinolytic activity was higher than those observed with BnP1. The differences of fibrino(geno)lytic activity between the SVMPs could contribute to the higher hemorrhagic action exhibited by jararhagin. However, this does not explain the disruption of the capillary vessels that is an essential requisite for blood extravasation. In this regard, a number of SVMPs from both P-I and P-III groups are reported to induce apoptosis of

endothelial cells. Proteolytic action was essential for apoptosis induced by HVI, a homodimeric P-III SVMP from Trimeresurus flavoviridis (Masuda et al., 2001), while the apoptosis induced by VLAIP, a P-III SVMP from Vipera libetina venom appears to be dependent on inhibition of adhesion endothelial cell to extracellular matrix proteins (Trummal et al., 2005). Apoptosis induced by halyase, another P-III from Cloydius halys venom (You et al., 2003) and BaP1, P-I isolated from B. asper is related to anoikis of endothelial cells (Dı´ az et al., 2005). In a recent publication, we reported that jararhagin is able to induce apoptosis by anoikis in Tend endothelial cell lines. This effect was completely dependent on catalytic activity and included a rapid change in cytoskeleton dynamics with cell retraction, accompanied by a rearrangement of actin network and a decrease in FAK association to actin and in tyrosine phosphorylated proteins, suggesting the interference with focal adhesion contacts (Tanjoni et al., 2005).

ARTICLE IN PRESS C. Baldo et al. / Toxicon 51 (2008) 54–65

64

1400

*

1200

CK (U/L)

1000

mechanisms involved in such activity, using more complex in vivo systems. Acknowledgments

800

We acknowledge the financial support by FAPESP, IOC-FIOCRUZ and CNPq.

600 400

References

200 0 PBS

BnP1 Jararhagin B. neu. venom SAMPLES

Fig. 7. Myotoxic effect of BnP1 and jararhagin in mice. Samples of 50 mg jararhagin and BnP1 or 10 mg B. neuwiedi venom were injected i.m. in Swiss mice (n ¼ 5). After 3 h, the animals were bled and the sera were assayed for creatine-kinase activity with a commercial kit. The data are expressed as mean7sd and significance was estimated by Student’s T test (* corresponds to po0.05, compared to PBS values).

In this study, we show that BnP1 and jararhagin were able to reduce viability of HUVECs also by interfering with attachment and inducing apoptosis. No significant differences were observed between the actions of jararhagin and BnP1, suggesting that the action of SVMPs on endothelial cells may participate but is not a unique condition for induction of hemorrhage. Concluding, the high plasticity of genes coding for SVMPs allowed during the evolution of snakes the generation of a number of structurally related molecules with functional variability, allowing better adaptation of snakes. This mechanism generated distinct proteolytic enzymes in venoms of the same species allowing the identification of two new members of P-I class of SVMPs in B. neuwiedi venom. BnP1 was very weakly hemorrhagic comparing the potent hemorrhagin jararhagin, a P-III SVMP. However, it presents activities similar to jararhagin towards fibrinogen and fibrin, and on endothelial cells, indicating that the presence of disintegrin-like and cysteine-rich domains is not fundamental for the expression of these biological activities of SVMPs. Moreover, the ability of SVMPs to interfere with individual components of the hemostatic systems appears to be insufficient to explain their differences in the hemorrhagic activity and further comparisons between BnP1 and jararhagin may be of great importance to explain the

Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller, W., Lipman, D.J., 1997. Gapped BLAST and PSIBLAST: a new generation of protein database search programs. Nucleic Acids Res. 25, 3389–3402. Bode, W., Gomis-Ruth, F., Stocker, W., 1993. Astacins, serralysins, snake venom and matrix metalloproteinases exhibit identical zinc-binding environments (HEXXH XXGXXH and met-turn) and topologies and should be grouped into a common family, the metzincins. FESBS Lett. 331, 134–140. Chippaux, J.P., Williams, V., White, J., 1991. Snake venom variability: methods of study, results and interpretation. Toxicon 29, 1279–1303. Clissa, P.B., Laing, G.D., Theakston, R.D.G., Mota, I., Taylor, M.J., Moura-da-Silva, A.M., 2001. The effect of jararhagin, a metalloproteinase from Bothrops jararaca venom, on proinflammatory cytokines released by murine peritoneal adherent cells. Toxicon 39, 1567–1573. Clissa, P.B., Lopes-Ferreira, M., Della-Casa, M.S., Farsky, S.H., Moura-da-Silva, A.M., 2006. Importance of jararhagin disintegrin-like and cysteine-rich domains in the early events of local inflammatory response. Toxicon 47, 591–600. Costa, E.P., Clissa, P.B., Teixeira, C.F.P., Moura-da-Silva, A.M., 2002. Importance of metalloproteinases and macrophages in viper snake envenomation-induced local inflammation. Inflammation 26, 13–17. Combet, C., Blanchet, C., Geourjon, C., Deleage, G., 2000. NPS@: Network Protein Sequence analysis. Trends Biochem. Sci. 25, 147–150. Dı´ az, C., Valverde, L., Brenes, O., Rucavado, A., Gutie´rrez, J.M., 2005. Characterization of events associated with apoptosis/anoikis induced by snake venom metalloproteinase BaP1 on human endothelial cells. J. Cell Biochem. 94, 520–528. Edman, P., Begg, G., 1967. A protein sequenator. Eur. J. Biochem. 1, 80–91. Fox, J.W., Serrano, S.M.T., 2005. Structural considerations of the snake venom metalloproteinases, key members of the M12 reprolysin family of metalloproteinases. Toxicon 45, 969–986. Gutie´rrez, J.M., Rucavado, A., 2000. Snake venom metalloproteinases: their role in the pathogenesis of local tissue damage. Biochimie 82, 841–850. Gutie´rrez, J.M., Rucavado, A., Escalante, T., Dı´ az, C., 2005. Hemorrhage induced by snake venom metalloproteinases: biochemical and biophysical mechanisms involved in microvessel damage. Toxicon 45, 997–1011. Ho, P.L., Serrano, S.M., Chudzinski-Tavassi, A.M., Moura da Silva, A.M., Mentele, R., Caldas, C., Oliva, M.L.C., Batista, I.F.C., Oliveira, M.L.S., 2002. Angiostatin-like molecules are

ARTICLE IN PRESS C. Baldo et al. / Toxicon 51 (2008) 54–65 generated by snake venom metalloproteinases. Biochem. Biophys. Res. Commun. 294, 879–885. Jaffe, E.A., Nachman, R.L., Becker, C.G., Minick, C.R., 1973. Culture of human endothelial cells derived from umbilical veins by morphologic and immunologic criteria. Identification. J. Clin. Invest. 52, 2745–2756. Jespersen, J., Astrup, T., 1983. A study of the fibrin plate assay of fibrinolytic agents: optimal conditions, reproducibility and precision. Haemostasis 13, 301–315. Kamiguti, A.S., Slupsky, J.R., Zuzel, M., Hay, C.R., 1994. Properties of fibrinogen cleaved by jararhagin, a metalloproteinase from the venom of Bothrops jararaca. Thromb. Haemost. 72, 244–249. Kamiguti, A.S., Hay, C.R.M., Theakston, R.D.G., Zuzel, M., 1996a. Insights into the mechanism of hemorrhage caused by snake venom metalloproteinases. Toxicon 34, 627–642. Kamiguti, A.S., Hay, C.R.M., Zuzel, M., 1996b. Inhibition of collagen-induced platelet-aggregation as the result of cleavage of a2b1-integrin by the snake venom metalloproteinase jararhagin. Biochem. J. 320, 635–641. Laemmli, U.K., 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680–685. Laing, G.D., Clissa, P.B., Theakston, R.D.G., Moura-da-Silva, A.M., Taylor, M.J., 2003. Inflammatory pathogenesis of snake venom metalloproteinase-induced skin necrosis. Eur. J. Immunol. 33, 3458–3463. Leon, I.R., Neves-Ferreira, A.G., Valente, R.H., Mota, E.M., Lenzi, H.L., Perales, J., 2007. Improved protein identification efficiency by mass spectrometry using N-terminal chemical derivatization of peptides from Angiostrongylus costaricensis, a nematode with unknown genome. J. Mass Spectrom. 42, 781–792. Masuda, S., Hayashi, H., Atoda, H., Morita, T., Araki, S., 2001. Purification, cDNA cloning and characterization of the vascular apoptosis-inducing protein, HV1, from Trimeresurus flavoviridis. Eur. J. Biochem. 268, 3339–3345. Modesto, J.C., Junqueira-de-Azevedo, I.L.M., Neves-Ferreira, A.G.C., Fritzen, M., Oliva, M.L.V., Ho, P.L., Perales, J., Chudzinski-Tavassi, A.M., 2005. Insularinase A, a prothrombin activator from Bothrops insularis venom, is a metalloprotease derived from a gene encoding protease and disintegrin domains. Biol. Chem. 386, 586–600. Moura-da-Silva, A.M., Theakston, R.D.G., Crampton, J.M., 1996. Evolution of disintegrin cystein-rich and mammalian matrix-degrading metalloproteinases: gene duplication and divergence of a common ancestor rather than convergent evolution. J. Mol. Evol. 43, 263–269. Moura-da-Silva, A.M., Della-Casa, M.S., David, A.S., Assakura, M.T., Butera, D., Lebrun, I., Shannon, J.D., Serrano, S.M.T.,

65

Fox, J.W., 2003. Evidences for heterogeneous forms of the snake venom metalloproteinase jararhagin: a factor contributing to snake venom variability. Arch. Biochem. Biophys. 490, 395–401. Moura-da-Silva, A.M., Butera, D., Tanjoni, I., 2007. Importance of snake venom metalloproteinases in cell biology: effects on platelets, inflammatory and endothelial cells. Curr. Pharmaceut. Des., in press. Nicolleti, I., Migliorati, G., Pagliacci, M.C., Gridnani, F., Riccardi, C., 1991. A rapid and simple method for measuring thymocyte apoptosis by propidium iodide staining and flow cytometry. J. Immunol. Methods 139, 271–279. Paine, M.J.I., Desmond, H.P., Theakston, R.D.G., Cramton, J.M., 1992. Purification, cloning, and molecular characterization of a high molecular weight hemorrhagic metalloprotease, jararhagin, from Bothrops jararaca venom. J. Biol. Chem. 267, 22869–22876. Rodrigues, V.M., Soares, A.M., Guerra-Sa´, R., Rodrigues, V., Fontes, M.R., Giglio, J.R., 2000. Structural and functional characterization of neuwiedase, a non-hemorrhagic fibrin (ogen)olytic metalloprotease from Bothrops neuwiedi snake venom. Arch. Biochem. Biophys. 381, 213–224. Rodrigues, V.M., Soares, A.M., Andria˜o-Escarso, S.H., Franceschi, A.M., Rucavado, A., Gutie´rrez, J.M., Giglio, J.R., 2001. Pathological alterations induced by neuwiedase, a metalloproteinase isolated from Bothrops neuwiedi snake venom. Biochimie 83, 471–479. Stocker, W., Bode, W., 1995. Structural features of a superfamily of zinc-endopeptidases: the metzincins. Curr. Opin. Struct. Biol. 5, 383–390. Swenson, S., Markland Jr., F.S., 2005. Snake venom fibrin (ogen)olytic enzymes. Toxicon 45, 1–19. Tanjoni, I., Weinlich, R., Della Casa, M.S., Clissa, P.B., Saldanha-Gama, R.F., Freitas, M.S., Barja-Fidalgo, C., Amarante-Mendes, G.P., Moura-da-Silva, A.M., 2005. Jararhagin, a snake venom toxin, induces a specialized form of apoptosis (anoikis) selective to endothelial cells. Apoptosis 10, 851–861. Trummal, K., Tonismagi, K., Siigur, E., Aaspollu, A., Lopp, A., Sillat, T., Saat, R., Kasak, L., Tammiste, I., Kogerman, P., Kalkkinen, N., Siigur, J, 2005. A novel metalloprotease from Vipera lebetina venom induces human endothelial cell apoptosis. Toxicon 46, 46–61. White, J., 2005. Snake venoms and coagulopathy. Toxicon 45, 951–967. You, W.K., Seo, H.J., Chung, K., Kim, D., 2003. A novel metalloproteinase from Cloydius halys venom induces endothelial cell apoptosis through its protease and disintegrinlike domains. J. Biochem. 134, 739–749.