at SciVerse ScienceDirect

Biochimie 95 (2013) 709e718

Contents lists available

Biochimie

journal homepage: www.elsevier .com/locate/biochi

Research paper

Purification and characterization of AHPM, a novel non-hemorrhagicP-IIIc metalloproteinase with a-fibrinogenolytic and plateletaggregation-inhibition activities, from Agkistrodon halys pallas venom

Jiajia Song, Xiaolong Xu*, Yan Zhang, Mingchun Guo, Xincheng Yan, Shasha Wang,Shang GaoDepartment of Chemistry, University of Science and Technology of China, No. 96, Jinzhai Road, Hefei 230026, PR China

a r t i c l e i n f o

Article history:Received 12 May 2012Accepted 17 October 2012Available online 24 October 2012

Keywords:Agkistrodon halys pallasNon-hemorrhagic metalloproteinaseAHPMa-FibrinogenasePlatelet aggregationZn2þ

Abbreviations: AHPM, metalloproteinase from thepallas; Apo-AHPM, metal-free AHPM; EDTA, ethyleSVMPs, snake venom metalloproteinases; PRP, platelpoor plasma; NNePF3, metalloproteinase from themetalloproteinase from the venom of Bothrops leloproteinase from the venom of Bothrops jararaloproteinase from the venom of Agkistrodonphenylmethanesulfonyl fluoride; pI, isoelectric poinfocusing; TCEP, tris-(2-carboxyethyl)phosphine hydroinhibitory concentration of platelet aggregation; ICplasma atomic emission spectrometer.* Corresponding author. Tel.: þ86 551 3603214; fax

E-mail address: xuxl@ustc.edu.cn (X. Xu).

0300-9084/$ e see front matter � 2012 Elsevier Mashttp://dx.doi.org/10.1016/j.biochi.2012.10.013

a b s t r a c t

A novel non-hemorrhagic metalloproteinase, AHPM, was purified from the venom of Agkistrodon halyspallas by a combination of ion-exchange and gel filtration chromatography. AHPM is a dimeric glyco-protein with multiple pIs around pH 7.9 and has a molecular mass of 110 kDa with two blocked N-terminuses. Partial sequence of AHPM obtained by LC-MS/MS analysis together with its dimeric naturereveals that it is a P-IIIc snake venom metalloproteinase composed of metalloproteinase, disintegrin-likeand cysteine-rich domains. AHPM has a conserved DECD sequence in the disintegrin-like domain. AHPMhydrolyzes casein and fibrinogen and also dissolves fibrin clots and the proteolytic activity is abolishedby EDTA, but not by PMSF, suggesting that it is a metalloproteinase. The protease hydrolyzes rapidly theAa-chain of fibrinogen followed by the Bb-chain and does not cleave the g-chain. AHPM containsendogenous Zn2þ and Ca2þ ions at a molar ratio of 1:1.9 and 1:4.2, respectively, and Zn2þ ions areessential for its proteolytic activity. AHPM inhibits collagen-and ADP-induced platelet aggregation withhalf maximal inhibitory concentrations of 200 � 8 nM and 280 � 10 nM, respectively. EDTA markedlyattenuates the inhibition of ADP-induced platelet aggregation by AHPM, indicating that the fibrinoge-nolytic activity of AHPM is involved in its inhibition of ADP-induced platelet aggregation. AHPM is devoidof hemorrhagic activity when injected (up to 30 mg) subcutaneously into mice. AHPM is so far identifiedas first non-hemorrhagic P-IIIc SVMP which has both fibrinolytic and platelet aggregation-inhibitionactivities. The bifunctional enzyme may have a potential clinical application as a thrombolytic agent.

� 2012 Elsevier Masson SAS. All rights reserved.

1. Introduction

Snake venoms are rich in a large variety of proteins and enzymes[1]. The venoms from Viperidae and Crotalidae contain abundantproteases that affect the blood coagulation cascade through bothpro-and anti-coagulant mechanisms [2]. Among them snake

venom of Agkistrodon halysne diaminetetraacetic acid;et-rich plasma; PPP, platelet-venom of Naja naja; Leuc-a,ucurus; BjussuMP-II, metal-cussu; Protease L4, metal-halys brevicaudus; PMSF,t; CIEF, capillary isoelectricchloride; IC50, half maximalPeAES, inductively coupled

: þ86 551 3603388.

son SAS. All rights reserved.

venom metalloproteinases (SVMPs) are of special research interestdue to their potential application as thrombolytic agents in thera-peutics and as biochemical tools in coagulation research anddiagnosis [3,4]. SVMPs are members of the Reprolysin subfamily ofthe M12 family of metalloproteinases and have been classified intothree groups according to their domain organization [5]. The P-Iclass only contains a single metalloproteinase domain, while the P-II class consists of a metalloproteinase domain and a disintegrindomain. The P-III class comprises metalloproteinase, disintegrin-like and cysteine-rich domains. Based on the position of theseventh cysteinyl residue in the metalloproteinase domain, P-IIImembers are organized into four subclasses [6]: P-IIIa, which isresistant to proteolytic processing; P-IIIb, which undergoes auto-proteolysis to release a 30 kDa fragment with disintegrin-andcysteine-rich domains; and P-IIIc, which forms a homodimericstructure [5]. The previous P-IV class containing two additionallectin-like domains compared with P-III SVMPs, has been reclassi-fied as P-IIId subclass. SVMPs display a wide array of biologicalactivities, most of which induce hemorrhage, such as alsophinase

J. Song et al. / Biochimie 95 (2013) 709e718710

from Alsophis portoricensis venom [7] and VLH2 from Vipera lebetinavenom [8]. Generally, the high molecular mass P-III SVMP, such asJararhagin from Bothrops jararaca venom is more strongly hemor-rhagic than the lower molecular mass P-I and P-II SVMPs [9].

Fibrinogen and platelets are basic components of thrombus thatplay an important role in its formation. The non-hemorrhagicSVMPs with both fibrinolytic activity and inhibitory activity onplatelet aggregation have the most promising application in drugdevelopment to treat thrombotic disorders [10]. However, manySVMPs with both fibrinolytic activity and inhibitory activity onplatelet aggregation are hemorrhagic, such as jararhagin [11],patagonfibrase [12] and atrolysin A [13]. On the other hand, manynon-hemorrhagic SVMPs with fibrinolytic activity are devoid ofinhibitory activity on platelet aggregation, such as fibrolase, a P-ISVMP obtained from Agkistrodon contortrix contortrix. Fibrolase notonly acts directly on fibrinogen but also degrades fibrin clots andhas been considered a potential therapeutic agent due to its veryeffective thrombolytic activity [14]. Currently, only a few non-hemorrhagic SVMPs with fibrinolytic and platelet aggregation-inhibition activities have been characterized from the venoms ofViperidae and Crotalidae, such as NNePF3 [15], and Leuc-a [16].

Agkistrodon halys pallas is the most venomous snake found inChina [17]. The venom of A. halys pallas contains a variety ofproteins and enzymes that disturb blood coagulation and fibrino-lytic systems [18,19]. Previously, a disintegrin-like and cysteine-richprotein named halysetin has been isolated from the venom. Haly-setin inhibits collagen-induced platelet aggregation, but does notinhibit ADP-stimulated platelet aggregation [20]. The present studyis the first report on the purification and characterization of a novelbasic a-fibrinogenolytic metalloproteinase, named AHPM, from thevenom of A. halys pallas. AHPM is a non-hemorrhagic P-IIIc SVMPwith both fibrinolytic and platelet aggregation-inhibition activitiesand may have a potential clinical application as a thrombolyticagent.

2. Materials and methods

2.1. Materials

Lyophilized A. halys pallas venom was purchased from theRainbow Snake Venom Company (Jiangxi, China). Mice weighted18e22 g were obtained from the Animal Services Center of AnhuiMedical University. All animal procedures were performed incompliance with the ethics committee on animal welfare of theUniversity of Science and Technology of China. Bovine fibrinogen,bovine thrombin, casein, Coomassie Brilliant Blue R250 and ADPwere purchased from SigmaeAldrich (St Louis, MO, USA). Collagenwas purchased from Helena laboratories (Beaumont, TX, USA).DEAE-Sephadex A-50 and Sephadex G-75 were purchased fromPharmacia Biotech (Uppsala, Sweden). Chelex-100 was purchasedfrom Bio-Rad Laboratories (Richmond, CA, USA). All other reagentsused were of analytical grade from commercial sources.

2.2. Purification of the metalloproteinase

Crude venom (3 g) was dissolved in 10 ml buffer (20 mM Tris-HCl, pH 8.0) and centrifuged to remove the precipitate. Thesupernatant was then applied to DEAE-Sephadex A-50(3.6 � 100 cm) which was previously equilibrated with the samebuffer and then elutedwith a linear gradient from0 to 0.5MNaCl inthe above buffer at a flow rate of 0.6 ml min�1. The fractions wereassayed for proteolytic activity using casein as a substrate (seebelow). The fraction with high proteolytic activity was pooled andsubsequently loaded onto a Sephadex G-75 column (2.6 � 100 cm),pre-equilibrated with 0.15 M NaCl in 20 mM Tris-HCl, pH 7.4 and

eluted with the same buffer. The fraction with proteolytic activitywas pooled, dialyzed against buffer A (20 mM Tris-HCl, pH 8.0) andthen applied to an SP-Sepharose High Performance column(16 � 25 mm), pre-equilibrated with buffer A. The proteins wereeluted first with buffer A (20 ml) and then with 0e100% increasinglinear gradient of buffer B (0.5 M NaCl in 20 mM Tris-HCl, pH 6.0)(90 ml). The main fraction with proteolytic activity was pooled,concentrated, dialyzed extensively against Milli-Q purified water,and then lyophilized with FD-1-50 freeze dryer (Beijing BoyikangInstruments Corporation, P.R. China). All the elution profiles weremonitored at 280 nm and the entire procedure was performed at4 �C. Metal-free AHPM (apo-AHPM) was prepared by dialysis ofpurified AHPM against a suspension of Chelex-100 (1 g/L) in 10 mMTris-HCl (pH 7.4).

2.3. SDS-polyacrylamide gel electrophoresis

The purity and molecular weight of the enzyme were analyzedby conventional SDS-PAGE. SDS-PAGE was performed at roomtemperature using 5% stacking and 12% resolving polyacrylamidegel under reducing and nonreducing conditions, and stained withCoomassie Brilliant Blue R250.

2.4. Capillary isoelectric focusing

The isoelectric point (pI) value was determined on P/ACE MDQcapillary isoelectric focusing (CIEF) (Beckman, USA) using carrierampholytes [21]. The sample was monitored at 280 nmwith a UVeVis detector (Beckman, USA) at 25 �C. A 25 kV constant voltage wasused during the whole focusing. Calibration standards (Beckman,USA) were between pH 3.59 and 9.46.

2.5. Determination of absorption coefficient at 280 nm

The lyophilized AHPM (4e5 mg) was accurately weighted by anAL104 electronic balance (Mettler-Toledo, China) and dissolved in20 mM Tris-HCl (pH 7.4) buffer. A UV-2100 spectrophotometer(Shimadzu, Japan) was used for the measurement of the ultravioletabsorption spectra of the enzyme. The experiments were per-formed three times and its absorption coefficients determined at280 nm were reproducible.

2.6. Analysis of carbohydrate content

Total carbohydrate was estimated by anthrone colorimetricassay [21]. 0.2 g of anthrone was dissolved in 100 ml of 98% (w/v)H2SO4. The anthrone reagent was freshly prepared just before use.1.0 ml of 1 mg/ml AHPM was added slowly to 3 ml of anthronereagent, which was placed in cold water to prevent excessiveheating. The mixture was heated at 100 �C for exactly 10 min inboiling water, and then cooled rapidly to 0 �C in ice water. Theabsorbance of the mixture at 620 nm was monitored to estimatethe content of carbohydrate. A standard curve of seven differentglucose concentrations (0e75 mg/ml) prepared by serial dilutionwas used in the calculation of the content of carbohydrate.

2.7. Determination of metal ions

The contents of Zn2þ, Ca2þ, Cu2þ, Mn2þ, Ni2þ, Mg2þ, Fe2þ andCo2þ in the purified AHPM were measured by inductively coupledplasma atomic emission spectrometer (ICPeAES) (Perkin ElmerCorporation, USA).

J. Song et al. / Biochimie 95 (2013) 709e718 711

2.8. N-terminal amino acid sequence analysis

The purified AHPM (40 mg) was run on a 12% SDS-PAGE gelunder reducing conditions and transferred to a PVDF membrane.After briefly stained with ponceau, the protein band of about50 kDa was cut and subjected to a PPSQ-33A protein sequenator(Shimadzu, Japan) to determine the first twenty N-terminal aminoacids [22].

2.9. LC-MS/MS analysis

The purified AHPM (20 mg) was run on a 12% SDS-PAGE gelunder reducing conditions. After briefly stained with CoomassieBrilliant Blue R-250, the protein band of about 50 kDa was cut andtransferred to a tube and destained with 30% acetonitrile in 50 mMNH4CO3. Then it was reduced with 10 mM dithiothreitol in 50 mMNH4CO3, alkylated with 30 mM iodoacetamide in 50 mM NH4CO3,and digested in gel with 0.1 mg trypsin overnight. The resultingpeptides were extracted from the gel with 60% (v/v) acetonitrile/5%(v/v) formic acid and concentrated for LC-MS/MS analysis ona Thermo Finnigan� linear IT mass spectrometer (LTQ) equippedwith an ESI source. The resulting fragment spectra were searchedagainst the Agkistrodon NCBI protein database using the SEQUESTalgorithm on Bioworks software (version 3.3) for peptideidentification.

2.10. Caseinolytic activity

Caseinolytic activity was measured according to the method ofRodrigues et al. [23]. 0.1 ml of sample solution was incubated with0.5 ml of 1% casein in 20 mM Tris-HCl buffer (pH 7.4) at 37 �C. After2 h, 1.0 ml of 5% trichloroacetic acid was added to stop the reaction,and then the mixture was centrifuged at 1800 � g for 10 min. Theproteolytic activity was estimated by reading the absorbance of theclear supernatant at 280 nm. One unit of caseinolytic activity wasdefined as the amount of enzyme which produces an absorbanceincrease of 0.001 units per minute. The effects of metal ions (Ca2þ,Cu2þ, Co2þ, Fe3þ, Mg2þ, Mn2þ, Ni2þ, Sr2þ and Zn2þ) and inhibitors(ethylenediaminetetraacetic acid (EDTA) and benzamidine) (1 mMin all cases, final concentration) were tested by pre-incubatingAHPM with each metal ion or inhibitor for 15 min at 37 �C beforeadding the casein solution.

2.11. Fibrinogenolytic activity

Bovine fibrinogen (20 mg) was incubated with AHPM (1 mg) in20 mM Tris-HCl (pH 7.4) containing 0.15 M NaCl and 1 mM CaCl2 at37 �C [12]. At various time intervals, aliquots of 8 ml of reactionmixture were transferred to 8 ml SDS-PAGE sample buffer (50 mMTris-HCl, pH 6.8, 20% glycerol, 2% SDS, 5% b-mercaptoethanol) andincubated at 100 �C for 5 min. The samples were then analyzed by12% SDS-PAGE under reducing conditions. For inhibition studies,AHPM was incubated with the following inhibitors: 1 mM benza-midine, 2 mM EDTA, 1 mM tris-(2-carboxyethyl)phosphinehydrochloride (TCEP) and 1 mM phenylmethanesulfonyl fluoride(PMSF), respectively, for 15 min at 37 �C before incubation withbovine fibrinogen.

2.12. Fibrinolytic activity

Fibrinolytic activity analysis was conducted using a fibrin platemethod with a slight modification [16]. Fibrin plates were preparedby the addition of 1.0 ml thrombin solution (8 U/ml) in 50 mM Tris-HCl, pH 7.4 to10 ml 1% bovine fibrinogen in 50 mM Tris-HCl, pH 7.4containing 0.15MNaCl and 1mM CaCl2. A fibrin clot was formed on

a level surface. Sample solution (10 ml) was added over the fibrinfilm and incubated at 37 �C for 18 h, and then the areas of clearancewere analyzed. Crude A. halys pallas venom (5 mg) was used aspositive control. The assay was also carried out in the presence ofmetalloproteinase inhibitor, EDTA (2 mM).

2.13. Hemorrhagic activity

Hemorrhagic activity was estimated by the method of Kondoet al. [24] with some modifications. Different doses of crude venomor AHPM (10, 20, 30 mg dissolved in 100 ml of saline) were injectedsubcutaneously into the dorsal skin of mice. The mice were sacri-ficed after 2 h and the skin was removed to measure the diametersof the hemorrhagic spot. Saline solution was utilized as negativecontrol. This assay was carried out by duplicates.

2.14. Assay for platelet aggregation

The effect of AHPM on the aggregation of platelets-rich plasma(PRP) was tested as described previously [21]. The study wasapproved by the local medical ethics committee for blood donationin Hefei (China). Whole blood was provided by healthy donors whohad not taken any medication prior to sampling for at least twoweeks. Blood was diluted (1:10) with 3.8% sodium citrate. PRP wasobtained after centrifuging of the diluted blood at 200� g for 5 minat 4 �C. The remaining blood was centrifuged at 940 � g for 10 minat 4 �C to obtain platelet-poor plasma (PPP). The final platelet countin PRP was adjusted to 250 � 103/ml using PPP. The inhibitoryactivity of AHPM on platelet aggregation was evaluated by stimu-lation of platelets with 10 mM ADP or 4 mg/ml collagen. Plateletaggregation was recorded for 10 min at 37 �C using a Chrono-logaggregometer (540 VS, USA).

3. Results

3.1. Purification of AHPM

The A. halys pallas crude venom was separated first on a DEAE-Sephadex A-50 column and eleven protein peaks were obtained(Fig. 1A). Fractions were assayed for the caseinolytic activity andpeak 1 showed remarkable caseinolytic activity. This fraction wasthen subjected to gel filtration on Sephadex G-75, yielding twopeaks (Fig. 1B). The first fraction (peak 1) with caseinolytic activitywas finally separated by chromatography on SP-Sepharose, yieldingthree peaks (Fig. 1C). The main fraction (peak 2) with caseinolyticactivity was collected, desalted and lyophilized. This purifiedproteinwas named as AHPM. The average yield of AHPM from 3 g ofcrude venom was about 10 mg. When analyzed by SDS-PAGE(Fig. 1D), the purified AHPM displayed a single band of 110 kDaunder non-reducing condition and gave a single band of 50 kDaunder reducing condition, suggesting that the purified AHPM isfairly homogeneous and consists of two chains with the samemolecular weight connected by disulfide linkage(s). The enzymehydrolyzed casein at pH 7.4 with the specific activity of5094 � 25 U/mg (mean � SD, n ¼ 3), while the crude venom hasa specific activity of 3311 � 4 U/mg (mean � SD, n ¼ 3) toward tocasein. The purification factor of AHPM is 1.5 (Table 1).

3.2. Properties

Ultraviolet absorption spectrum of AHPM showed that thepurified AHPM absorbed strongly at 280 nm with an absorptioncoefficient (A1%

280) of 8.5. The content of carbohydrate in AHPMwas determined to be 2.5 � 0.1% (mean � SD, n ¼ 3) of the totalprotein by anthrone colorimetric assay, indicating that it is

Fig. 1. Purification of AHPM from Agkistrodon halys pallas venom. (A) Anion-exchange chromatography of the crude venom (3 g) on a DEAE-Sephadex A-50 column (3.6 � 100 cm).The column was washed with 20 mM Tris-HCl buffer (pH 8.0) at a flow rate of 0.6 ml min�1, and then eluted with a linear gradient of 0e0.5 M NaCl in the same buffer, at 4 �C. (B)Chromatography of the proteins obtained from the peak 1 in the first step on a Sephadex G-75 column (2.6 � 100 cm). The column was eluted with 0.15 M NaCl in 20 mM Tris-HCl(pH 7.4) at a flow rate of 0.25 ml min�1, at 4 �C. (C) Chromatography of the proteins obtained from the peak 1 in panel (B) on a SP-Sepharose column (16 � 25 mm). The column waseluted first with 20 ml of buffer A (20 mM Tris-HCl, pH 8.0) and then with 0e100% increasing linear gradient of 90 ml of buffer B (0.5 M NaCl in 20 mM Tris-HCl, pH 6.0) at a flowrate of 2 ml min�1, at 4 �C. All the profiles were monitored by absorbance at 280 nm (D) SDS-polyacrylamide gel electrophoresis of AHPM. Purified AHPM was electrophoresed in12% polyacrylamide gels containing 0.1% SDS under nonreducing (NR) and reducing (R) condition with 0.1 M b-mercaptoethanol. Lane 1 and lane 3, AHPM; lane 2, molecular massmarkers. The gel was stained in 0.25% Coomassie Blue R-250.

J. Song et al. / Biochimie 95 (2013) 709e718712

a glycoprotein. The molar ratios of Zn2þ and Ca2þ to the purifiedAHPM were determined to be 1.9 � 0.1 and 4.2 � 0.2 (mean � SD,n ¼ 3), respectively, and no Mg2þ, Fe2þ, Mn2þ, Cu2þ, Ni2þ and Co2þ

were detected in the purified AHPM by ICP-AES, indicating thateach purified AHPM molecule contains two Zn2þ and four Ca2þ.CIEF was used to determine the pI of AHPM. Characteristically fora glycoprotein, the purified AHPM showed one broad peak centeredat pH 7.9 in Fig. 2, suggesting that AHPM has multiple pIs aroundpH 7.9.

3.3. Amino acid sequence analysis

Purified AHPM was resistant to Edman degradation, indicatingthat both N-terminuses of AHPM are blocked. In further studies,

Table 1Purification of AHPM from Agkistrodon halys pallas venom.

Step Total protein(mg)

Totalactivity(units)

Specificactivity(units/mg)

Purification(fold)

Yield(%)

Crude venom 4000 1.3 � 106 3311 1.0 100DEAE-Sephadex

A-50197 2.0 � 105 1035 0.3 4.9

Sephadex G-75 31 1.4 � 105 4410 1.3 0.8SP-Sepharose 10 5.1 � 104 5094 1.5 0.25

purified AHPM was digested in gel with trypsin and the resultingpeptides were analyzed by LC-MS/MS. The result of SEQUESTsearching of the LC-MS/MS spectra of the trypsin-digested frag-ments showed that there are 8 fragments, totally 151 amino acidresidues, matched with halysase, a P-IIIa SVMP from Gloydius halysvenom of Korea [25]. As shown in Fig. 3, three fragments (P1, P2 andP3) are located in themetalloproteinase domain, two fragments (P4

Fig. 2. Capillary isoelectric focusing profile of AHPM and pI markers. The pI of AHPMwas estimated according to the pI markers: ribonuclease A (9.46), carbonic anhydraseII (5.77), b-lactoglobulin A (5.23) and CCK flanking peptide (3.59).

Fig. 3. The partial amino acid sequence of AHPM obtained by LC-MS/MS analysis and aligned with sequences of the most homologous P-III SVMPs: halysase (the UniProtKBaccession number is Q8AWI5), agkihagin from Agkistrodon acutus (Q1PS45), VAP1 (Q9DGB9), TSV-DM (Q2LD49) and halysetin (Q90Y44). The peptides (P1-P8) of AHPM identified byLC-MS/MS and the identical amino acid residues are gray-backgrounded. The metalloproteinase, disintegrin-like and cysteine-rich domains are indicated by asterisk, dash and dot,respectively. The putative a2b1 integrin binding sequences, DECD, are gray-backgrounded and highlighted.

J. Song et al. / Biochimie 95 (2013) 709e718 713

Fig. 5. Fibrinogenolytic activity of AHPM. (A) Bovine fibrinogen was incubated withAHPM for different times in 20 mM Tris-HCl, pH 7.4 containing 0.15 M NaCl and 1 mMCaCl2 and then checked by SDS-PAGE (12% gel) under reducing conditions. Lanes 1e11,fibrinogen incubated with AHPM for 0, 5, 15, 30, 60, 120, 180, 240, 300, 360 and420 min, respectively. Lane 12, molecular mass markers. (B) Effect of inhibitors on thedigestion of fibrinogen by AHPM, analyzed by SDS-PAGE (12% gel), under reducingconditions. Lane 1, the control of bovine fibrinogen incubated at 37 �C for 1 h with noAHPM. Lanes 2e8, AHPM was incubated with buffer, 2 mM EDTA, 1 mM TCEP, 1 mMPMSF, 1 mM benzamidine, 2 mM 1,10-phenantroline and 2 mM EGTD for 15 min at37 �C, respectively, before the incubation with bovine fibrinogen at 37 �C for 1 h.

J. Song et al. / Biochimie 95 (2013) 709e718714

and P5) in disintegrin-like domain, three fragments (P6, P7 and P8)in the cysteine-rich domain, classifying AHPM into P-III class ofSVMPs. Since AHMP exists as a dimer, we can classify it in the P-IIIcsubclass. All of the identified AHPM peptides exhibit a high degreeof sequence identity with other P-III SVMPs.

3.4. Effects of metal ions and inhibitors on caseinolytic activity

The effects of metal ions and proteinase inhibitors on thecaseinolytic activity of AHPMwere investigated. As shown in Fig. 4,the caseinolytic activity of AHPM was little affected by Ca2þ andSr2þ, and slightly inhibited by Mg2þ and Cu 2þ, but markedlyinhibited by Mn2þ, Co2þ and Zn2þ. Ni2þ and Fe3þ inhibited almostall caseinolytic activity. AHPM was significantly inhibited by EDTA(93%), but not inhibited by benzamidine, an inhibitor of serineproteinase, indicating that AHPM is ametalloproteinase. AHPMwasalso significantly inhibited by TCEP (95%), revealing that thedisulfide bonds play a critical role in the caseinolytic activity. Apo-AHPM could recover 97% and 9% of the caseinolytic activity,respectively, in the presence of 5 mMZn2þ and 5 mMCa2þ, indicatingthat Zn2þ is essential for its caseinolytic activity.

3.5. Fibrinogenolytic and fibrinolytic activities

The fibrinogenolytic activity of AHPM was analyzed by SDS-PAGE. As shown in Fig. 5A, AHPM degraded the Aa-chain ofbovine fibrinogenmuch faster than the Bb-chains, while no obviouscleavage of the g-chain was detected, even after 7 h of incubation.The degradation of the Aa chains occurred between 5 and 30 min.In contrast, the degradation of the Bb chains was between 1 and 5 h.Based on these characteristics, AHPM was classified as an a-fibri-nogenase. As shown in Fig. 5B, the fibrinogenolytic activity ofAHPM was completely inhibited by EDTA, but did not inhibited bybenzamidine and PMSF, further revealing that AHPM is a metal-loproteinase. TCEP also completely inhibited the fibrinogenolyticactivity of AHPM, suggesting that the disulfide bonds also play

Fig. 4. Effects of metal ions and inhibitors on the caseinolytic activity of AHPM. 10 mgAHPM was incubated with 0.1 ml of each metal ion or inhibitor or 10 mg apo-AHPMwas incubated with 0.1 ml of Ca2þ or Zn2þ in 20 mM Tris-HCl buffer (pH 7.4) for15 min at 37 �C and then mixed with 0.5 ml of 1% casein in same buffer (pH 7.4) andincubated at 37 �C. After 2 h, 1.0 ml of 5% trichloroacetic acid was added to stop thereaction, and the mixture was centrifuged at 1800 � g for 10 min. The proteolyticactivity was estimated by reading the absorbance of the clear supernatant at 280 nm.The purified AHPM was used as a positive control. Results shown are mean � SD, n ¼ 3.The final concentration of each metal ion or inhibitor is 1 mM for AHPM assay and thefinal concentration of Zn2þ or Ca2þ is 5 mM for apo-AHPM assay.

a critical role in the fibrinogenolytic activity. To analyze the roles ofZn2þ and Ca2þ in the activity of AHPM, the inhibitions of fibrino-genolytic activity of AHPM by 1,10-phenantroline and EGTA werealso analyzed by SDS-PAGE. 1,10-phenantroline and EGTA arespecific chelators of Zn2þ and Ca2þ, respectively [26,27]. As shownin Fig. 5B, the removal of Zn2þ from AHPM by 2 mM 1,10-phenantroline results in loss of the fibrinogenolytic activity ofAHPM, however, the removal of Ca2þ from AHPM by 2 mM EGTAdoes not obviously affect on the fibrinogenolytic activity of AHPM,which indicates that Zn2þ is essential for its fibrinogenolyticactivity.

Fibrinolytic activity was assayed using the fibrin plate method.As shown in Fig. 6, addition of 5 mg crude venom and 5 mg AHPM ledto the formation of a clear hollow of 123 � 8 and 44 � 5 mm2

(mean � SD, n ¼ 3), respectively. EDTA completely inhibited thefibrinolytic activity of AHPM, which further confirmed that AHPMis a metalloproteinase.

3.6. Hemorrhagic activity

As shown in Fig. 7, 3 mg of the crude venom induced a conspic-uous hemorrhage after subcutaneous injection in mice. In contrast,AHPM did not induce hemorrhage up to a dose of 30 mg, indicatingthat AHPM is a non-hemorrhagic metalloproteinase.

3.7. Platelet aggregation

Inhibition of platelet aggregation by AHPMwas examined using3-min pre-incubation of various concentrations of AHPM withplatelets before adding ADP or collagen as the inducers. As shown

Fig. 6. Fibrinolytic activity of AHPM. Fibrin plates were prepared by the addition of1.0 ml thrombin solution (8 U/ml) in 50 mM Tris-HCl, pH 7.4e10 ml 1% bovinefibrinogen in 50 mM Tris-HCl, pH 7.4 containing 0.15 M NaCl and 1 mM CaCl2. Thesamples were placed in the central of the plate and incubated for 18 h at 37 �C. (A)Saline (control); (B) Crude Agkistrodon halys pallas venom (5 mg); (C) AHPM (5 mg); (D)AHPM (5 mg) þ EDTA (2 mM).

J. Song et al. / Biochimie 95 (2013) 709e718 715

in Fig. 8, AHPM strongly inhibited the ADP- or collagen-inducedplatelet aggregation in a dose-dependent manner. The halfmaximal inhibitory concentration (IC50) values for ADP-andcollagen-stimulated platelet aggregation were 280 � 10 nM and200 � 8 nM (mean � SD, n ¼ 3), respectively, suggesting that theinhibition of ADP-induced platelet aggregation by AHPM is lesspotent than that of collagen-induced platelet aggregation. Whencollagen was pre-incubated with AHPM at 37 �C for 3 min, it couldinduce platelet aggregation in the same way as by pre-incubatingthe enzyme with platelet suspensions at 37 �C for 3 min (Fig. 8E),indicating that pre-incubation of collagen with AHPM does notchange its inhibitory effect on the collagen-induced plateletaggregation. As shown in Fig. 8F, EDTA markedly abolished theinhibition of platelet aggregation by AHPM, suggesting that metalions in AHPM play a key role in its platelet aggregation-inhibitionactivity. Similarly, TCEP also significantly abolished the inhibitionof platelet aggregation by AHPM, indicating that the disulfides

Fig. 7. Hemorrhagic activity of AHPM. One-hundred microliters of samples were subcutaneothe skin was removed to measure the hemorrhagic spot. (A) Saline (control); (B) Crude Agk

bonds are also essential for its platelet aggregation-inhibitionactivity.

4. Discussion

SVMPs are the major component in Viperidae and Crotalidaevenoms. Among three SVMP classes, the P-III SVMPs have multiplefunctions and may serve as excellent probes for clarifying themechanisms underlying thrombosis and hemostasis. Most P-IIISVMPs are capable of producing hemorrhage. Only a few P-IIISVMPs are not hemorrhagic and may have more potential appli-cations in dissolving thrombi than thrombin-like enzymes, whichonly diminish fibrinogen but do not affect thrombi [28]. In thiswork, a novel non-hemorrhagic P-IIIc metalloproteinase, namedAHPM, was purified from the A. halys pallas venom by a combina-tion of ion exchange and gel filtration chromatography.

AHPM is a dimeric glycoprotein with a molecular mass of110 kDa. Partial sequence of the protein obtained by LC-MS/MSanalysis reveals a sequence similarity with other members of P-IIIc SVMPs, VAP1 from the Crotalus atrox venom [29] and TSV-DMfrom the Trimeresurus stejnegeri venom [30] (Fig. 3). P-IIIc SVMPsusually undergo post-translational modifications to form a homo-dimer [5]. The two chains of AHPM have the same molecularweight as determined by SDS-PAGE (Fig. 1D) and the N-terminusesof both chains are blocked. These results taken together suggestthat like other P-IIIc SVMPs, AHPM may be a homo-dimer. Asmentioned previously, P-IIIb SVMPs can undergo autoproteolysis torelease a 30 kDa fragment with disintegrin-and cysteine-richdomains. To analyze whether AHPM can autoproteolyze underphysiological condition, 10 mg AHPM was incubated with 20 mMTris-HCl buffer (pH 7.4) at 37 �C for 24 h. After analysis of thereaction mixture by SDS-PAGE, no obvious fragments were found(data not shown), indicating the high stability of the enzyme andgetting rid of the possibility of P-IIIb SVMP.

AHPM and halysetin, a disintegrin-like/cysteins-rich protein,were isolated from the same venom of A. halys pallas. Partialsequence of AHPM identified by LC-MS/MS displays a highhomology to halysetin (58.1% identity), revealing that halysetin isnot an autoproteolysis product of AHPM. Despite the fact that thecomplete primary structure of AHPM is not known yet, it is the firstdimeric SVMP identified in the A. halys pallas venom.

Recently, the crystal structures of seven P-III SVMPs, such asVAP1, a P-IIIc, VAP2B, a P-IIIb, both from the C. atrox venom, and

usly injected in the mice. After 2 h, animals were sacrificed and the injected portion ofistrodon halys pallas venom (3 mg); (C) AHPM (30 mg).

Fig. 8. Inhibition of ADP-and collagen-induced platelet aggregation by AHPM. (A) Platelet-rich plasma (PRP) was preincubated with various concentrations of AHPM (0e600 nM)for 3 min prior to the stimulation of platelet aggregation by 10 mM ADP. (B) Concentration dependence of AHPM inhibition of ADP-induced human platelet aggregation. (C) PRP waspreincubated with various concentrations of AHPM (0e600 nM) for 3 min prior to the stimulation of platelet aggregation by 4 mg/ml collagen. (D) Concentration dependence ofAHPM inhibition of collagen-induced human platelet aggregation. (E) PRP was preincubated with 200 nM AHPM for 3 min prior to the stimulation of platelet aggregation by 4 mg/mlcollagen (1) and 4 mg/ml collagen was preincubated with 200 nM AHPM for 3 min and then incubated with PRP (2). (F) PRP was preincubated with saline (control) (1) and 200 nMAHPM in the absence (4) and presence of 400 mM TCEP (2) or 400 mM EDTA (3) for 3 min prior to the stimulation of platelet aggregation by 10 mM ADP. Results shown aremean � SD, n ¼ 3.

J. Song et al. / Biochimie 95 (2013) 709e718716

RVV-X, a P-IIId from the Vipera russelli venom, have been deter-mined [5]. All the seven P-III SVMPs share high homologoussequences with very similar backbone structures. They all containone Zn2þ and three Ca2þ ions, except for VAP1 that contains twoZn2þ ions and four Ca2þ ions. The purified AHPM contains 1.9 molendogenous Zn2þ and 4.2 mol endogenous Ca2þ per mole ofmolecule, indicating that like VAP1, each AHPM molecule containstwo Zn2þ ions and four Ca2þ ions. The purified AHPM still bindswith two Zn2þ ions and four Ca2þ ions after the purification process,suggesting that AHPM has a high binding affinity for Zn2þ and Ca2þ.Ca2þ ions are usually important for the stabilization of SVMPs [16].Although Zn2þ ions are essential for the proteolytic activity of

AHPM, they are also able to inhibit its proteolytic activity at highconcentration (1 mM) (Fig. 4), as observed for Leuc-a [16]. Apossible explanation is that besides the two strong Zn2þ-bindingsites, AHPM may have weak Zn2þ-binding site(s), and the bindingof Zn2þ in the weak site(s) probably inhibits its proteolytic activity.Further investigation is necessary to clarify the complex role ofZn2þ in the proteolytic activity of AHPM. AHPM is an alkalinemetalloproteinase with multiple pIs around pH 7.9, which is similarto VAP1 (pI ¼ 7.5) [29] and bothrojaractivase from B. jararacavenom (pI ¼ 8.1) [31].

AHPM possesses proteolytic activity toward casein and fibrin-ogen. Based on its ability to preferentially attack the Aa chain of

J. Song et al. / Biochimie 95 (2013) 709e718 717

fibrinogen, AHPM is classified as an a-fibrinogenase. AHPM notonly acts directly on fibrinogen but also degrades fibrin clots. Unlikemost P-III SVMPs that are characterized by higher hemorrhagic,AHPM is devoid of hemorrhagic activity, suggesting that theenzyme does not degrades the basement membrane componentsof capillary vessels. AHPM also does not cleave Factor IX and FactorX (data not shown). AHPM inhibits the collagen-and ADP-inducedplatelet aggregation in the order collagen > ADP, as observed forNNePF3 [15]. The IC50 value for AHPM (200 � 8 nM) is about 3times less active than NNePF3 (IC50 ¼ 75 � 5 nM) in blockingcollagen-induced platelet aggregation, and the IC50 value for AHPM(280 � 10 nM) is about 1.5 times less active than NNePF3(IC50 ¼ 185 � 10 nM) and 14 times more active than Leuc-a (IC50¼ 2.8 mM) in blocking ADP-induced platelet aggregation [16].

The inhibition of ADP-induced platelet aggregation by AHPMmay be the consequence of degradation of fibrinogen by theenzyme, since fibrinogen is indispensable to cross-link the plateletsthrough the membrane glycoprotein IIb/IIIa (integrin aIIbb3) [32].The Aa chain of fibrinogen contains two ArgeGlyeAsp (RGD)sequences that recognize and bind to integrin aIIbb3 [33]. AHPMattacks on the Aa chain and releases RGD-containing peptideswhich act as the competitive inhibitor for the binding betweenfibrinogen and integrin aIIbb3 and enhance the inhibition of plateletaggregation by AHPM. EDTA markedly abolishes the inhibition ofplatelet aggregation by AHPM, which confirms that the fibrinoge-nolytic activity of AHPM plays an important role in its inhibition ofADP-induced platelet aggregation. The extent of ADP-inducedplatelet aggregation in the presence of AHPM rapidly increases toreach a maximum and then reverses, leading to obvious disaggre-gation (Fig. 8A). The disaggregation may be attributed to thecompetitive inhibition by RGD-containing peptides from the Aachain of fibrinogen, which further suggests the role of fibrinoge-nolytic activity of AHPM in the aggregation inhibition process.

Interestingly, the shape of AHPM-inhibition curve for collagen-induced platelet aggregation was different from that for ADP-induced platelet aggregation. The extent of collagen-inducedplatelet aggregation in the presence of AHPM was graduallyincreasing and reached a plateau and no disaggregation wasobserved (Fig. 8C), suggesting that AHPM inhibits collagen-andADP-induced platelet aggregation probably through differentmechanisms. The major collagen receptors on platelet membraneare integrin a2b1 and GPVI [34]. The collagen-induced activationand aggregation of platelets involve complex mechanismsinvolving initial adhesion of platelets to collagen through theintegrin a2b1 and subsequent activation of intracellular signalingmechanisms through GPVI [15]. The inhibition of collagen-inducedplatelet aggregation by SVMPs is usually attributed to the interac-tion of SVMPs with collagen, integrin a2b1 or GPVI as well as thedegradation of collagen, integrin a2b1 or GPVI by SVMPs [28,35,36].As shown in Fig. 8E, both collagen pre-incubated with AHPM andcollagen added after incubation of platelet suspensions with AHPMcause similar values of platelet aggregation, suggesting that theinhibitory activity on collagen-induced aggregation may resultfrom the interaction of AHPM with platelet collagen receptorswithout affecting collagen itself. It has been reported that the dis-integrins from snake venoms contain the characteristic RGDsequence and bind to integrin aIIbb3 through RGD sequence andprevent the binding of fibrinogen to integrin aIIbb3 in the aggre-gation process and thus inhibit platelet aggregation [15]. AHPM hasa DECD sequence in the disintegrin-like domain which is ina homologous position to the RGD motif of the disintegrins (Fig. 3).DECD motif is thought to recognize the collagen receptor a2b1integrin on platelet membrane [25]. Therefore, the inhibitioncollagen-induced aggregation by AHPM is likely attributed to theinteraction AHPM with integrin a2b1 through the DECD motif, as

observed for NNePF3 [15]. Further investigation is necessary toelucidate this issue.

Currently, four non-hemorrhagic SVMPs, Leuc-a [16,37],BjussuMP-II [38], protease L4 [32] and NNePF3 [15], have beenreported to possess fibrinolytic and platelet aggregation-inhibitionactivities. Leuc-a, BjussuMP-II and protease L4 are P-I SVMPs, whileNNePF3 is a P-IIIa SVMP. To the best of our knowledge, AHPM is theone and only non-hemorrhagic P-IIIc SVMP which has both fibri-nolytic and platelet aggregation-inhibition activities. Its bifunc-tional activity, combined with its non-hemorrhagic activity, makesAHPM a promising thrombolytic agent and an interesting tool incoagulation research and diagnosis.

5. Conclusion

In this work, we have purified and characterized a novel non-hemorrhagic metalloproteinase, AHPM, from A. halys pallasvenom. AHPM is a dimeric glycoprotein containing two Zn2þ andfour Ca2þ and possesses both fibrinolytic and platelet aggregation-inhibition activities. Based on its molecular mass and partial aminosequence, AHPM is a P-IIIc SVMP, composed of metalloproteinase,disintegrin-like and cysteine-rich domains. The finding providesa new resource for developing therapeutic agents to prevent andtreat thrombosis. Further investigations are under way to deter-mine the precise mechanisms of the degradation of fibrin and theinhibition of platelet aggregation by the enzyme.

Acknowledgments

This work was supported by grants from the National NaturalScience Foundation of China (Grant No. 21171157, 20871111,20571069).

References

[1] F.S. Markland, Snake venoms and the hemostatic system, Toxicon 36 (1998)1749e1800.

[2] R.B. Zingali, Interaction of snake-venom proteins with blood coagulationfactors: mechanisms of anticoagulant activity, Toxin. Rev. 26 (2007) 25e46.

[3] S. Moll, P. Kenyon, L. Bertoli, J. De Maio, H. Homesley, S.R. Deitcher, Phase IItrial of alfimeprase, a novel-acting fibrin degradation agent, for occludedcentral venous access devices, J. Clin. Oncol. 24 (2006) 3056e3060.

[4] S. Braud, C. Bon, A. Wisner, Snake venom proteins acting on hemostasis,Biochimie 82 (2000) 851e859.

[5] S. Takeda, H. Takeya, S. Iwanaga, Snake venom metalloproteinases: structure,function and relevance to the mammalian ADAM/ADAMTS family proteins,Biochim. Biophys. Acta 1824 (2012) 164e176.

[6] J.W. Fox, S.M.T. Serrano, Insights into and speculations about snake venommetalloproteinase (SVMP) synthesis, folding and disulfide bond formation andtheir contribution to venom complexity, FEBS J. 275 (2008) 3016e3030.

[7] C.L. Weldon, S.P. Mackessy, Alsophinase, a new P-III metalloproteinase with a-fibrinogenolytic and hemorrhagic activity from the venom of the rear-fangedPuerto Rican Racer Alsophis portoricensis (Serpentes: Dipsadidae), Biochimie94 (2012) 1189e1198.

[8] L. Hamza, C. Gargioli, S. Castelli, S. Rufini, F. Laraba-Djebari, Purification andcharacterization of a fibrinogenolytic and hemorrhagic metalloproteinaseisolated from Vipera lebetina venom, Biochimie 92 (2010) 797e805.

[9] A.M. Moura-da-Silva, O.H.P. Ramos, C. Baldo, S. Niland, U. Hansen, J.S. Ventura,S. Furlan, D. Butera, M.S. Della-Casa, I. Tanjoni, P.B. Clissa, I. Fernandes,A.M. Chudzinski-Tavassi, J.A. Eble, Collagen binding is a key factor for thehemorrhagic activity of snake venom metalloproteinases, Biochimie 90 (2008)484e492.

[10] S. Swenson, L.R. Bush, F.S. Markland, Chimeric derivative of fibrolase, a fibri-nolytic enzyme from southern copperhead venom, possesses inhibitoryactivity on platelet aggregation, Arch. Biochem. Biophys. 384 (2000) 227e237.

[11] L.J. Lambert, A.A. Bobkov, J.W. Smith, F.M. Marassi, Competitive interactions ofcollagen and a jararhagin-derived disintegrin peptide with the integrin a2-Idomain, J. Biol. Chem. 283 (2008) 16665e16672.

[12] M.E. Peichoto, P. Teibler, S.P. Mackessy, L. Leiva, O. Acosta, L.R. Gonçalves,A.M. Tanaka-Azevedo, M.L. Santoro, Purification and characterization ofpatagonfibrase, a metalloproteinase showing a-fibrinogenolytic and hemor-rhagic activities, from Philodryas patagoniensis snake venom, Biochim. Bio-phys. Acta 1770 (2007) 810e819.

J. Song et al. / Biochimie 95 (2013) 709e718718

[13] A.F. Pinto, L. Ma, B. Dragulev, J.A. Guimarães, J.W. Fox, Use of SILAC forexploring sheddase and matrix degradation of fibroblasts in culture by the PIIISVMP atrolysin A: identification of two novel substrates with functionalrelevance, Arch. Biochem. Biophys. 465 (2007) 11e15.

[14] S. Swenson, F.S. Markland, Fibrolase, Toxin. Rev. 25 (2006) 351e378.[15] M.S. Kumar, K.S. Girish, B.S. Vishwanath, K. Kemparaju, The metalloprotease,

NN-PF3 from Naja naja venom inhibits platelet aggregation primarily byaffecting a2b1 integrin, Ann. Hematol. 90 (2011) 569e577.

[16] C.A. Bello, A.L.N. Hermogenes, A. Magalhães, S.S. Veiga, L.H. Gremski,M. Richardson, E.F. Sanchez, Isolation and biochemical characterization ofa fibrinolytic proteinase from Bothrops leucurus (white-tailed jararaca) snakevenom, Biochimie 88 (2006) 189e200.

[17] K. Xu, b-Agkistrodotoxin - a Presynaptic Neurotoxin from venom of the PitViper (Agkistrodon-Blomhoffii-Brevicaudus Stejneger), Chin. Sci. Bull. 35 (1990)1145e1155.

[18] C.Y. Fan, Y.C. Qian, S.L. Yang, Y. Gong, Cloning, sequence analysis and expressionin E. coli of the cDNA of the thrombin-like enzyme (pallabin) from the venom ofAgkistrodon halys Pallas, Biochem. Mol. Biol. Int. 47 (1999) 217e225.

[19] H. Wu, X. Xu, D. Shen, L. Peng, J. Song, Y. Zhang, Binding of Ca2þ and Zn2þ tofactor IX/X-binding protein from venom of Agkistrodon halys Pallas: stabili-zation of the structure during GdnHCl-induced and thermally induceddenaturation, J. Biol. Inorg. Chem. 16 (2011) 69e79.

[20] J.W. Liu, X.Y. Du, P. Liu, X. Chen, J.M. Xu, X.F. Wu, Y.C. Zhou, Purification,characterization, and cDNA sequence of halysetin, a disintegrin-like/cysteine-rich protein from the venom of Agkistrodon halys Pallas, Biochem. Biophys.Res. Commun. 278 (2000) 112e118.

[21] L. Peng, X.L. Xu, D.K. Shen, Y. Zhang, J.J. Song, X.C. Yan, M.C. Guo, Purificationand partial characterization of a novel phosphodiesterase from the venom ofTrimeresurus stejnegeri: inhibition of platelet aggregation, Biochimie 93 (2011)1601e1609.

[22] S. Cherif, A. Fendri, N. Miled, H. Trabelsi, H. Mejdoub, Y. Gargouri, Crabdigestive lipase acting at high temperature: purification and biochemicalcharacterization, Biochimie 89 (2007) 1012e1018.

[23] V.M. Rodrigues, A.M. Soares, R. Guerra-Sá, V. Rodrigues, M.R.M. Fontes,J.R. Giglio, Structural and functional characterization of neuwiedase, a non-hemorrhagic fibrin(ogen)olytic metalloprotease from Bothrops neuwiedi snakevenom, Arch. Biochem. Biophys. 381 (2000) 213e224.

[24] H. Kondo, S. Kondo, H. Ikezawa, R. Murata, Studies on the quantitative methodfor determination of hemorrhagic activity of habu snake venom, Jpn. J. Med.Sci. Biol. 13 (1960) 43e52.

[25] W.K. You, H.J. Seo, K.H. Chung, D.S. Kim, A novel metalloprotease from Gloy-dius halys venom induces endothelial cell apoptosis through its protease anddisintegrin-like domains, J. Biochem. 134 (2003) 739e749.

[26] H. Venter, G. Osthoff, D. Litthauer, Purification and characterization of a met-alloprotease from chryseobacterium indologenes Ix9a and determination of

the amino acid specificity with electrospray mass spectrometry, Protein Expr.Purif. 15 (1999) 282e295.

[27] T. Azuma, S. Ikeda, T. Kondo, H. Imai, M. Yamada, Ethylenediamine-N, N, N0 ,N0-tetraacetic acid induces parthenogenetic activation of porcine oocytes atthe germinal vesicle stage, leading to formation of blastocysts, Biol. Reprod. 64(2001) 647e653.

[28] W.J. Wang, Purification and functional characterization of AAV1, a novel P-IIImetalloproteinase, from Formosan Agkistrodon acutus venom, Biochimie 89(2007) 105e115.

[29] S. Takeda, T. Igarashi, H. Mori, S. Araki, Crystal structures of VAP1 revealADAMs’ MDC domain architecture and its unique C-shaped scaffold, EMBO J.25 (2006) 2388e2396.

[30] S.G. Wan, Y. Jin, W.H. Lee, Y. Zhang, A snake venom metalloproteinase thatinhibited cell proliferation and induced morphological changes of ECV304cells, Toxicon 47 (2006) 480e489.

[31] M. Berger, A.F. Pinto, J.A. Guimarães, Purification and functional character-ization of bothrojaractivase, a prothrombin-activating metalloproteinaseisolated from Bothrops jararaca snake venom, Toxicon 51 (2008) 488e501.

[32] S. Fujimura, T. Rikimaru, S. Baba, J. Hori, X.Q. Hao, S. Terada, E. Kimoto,Purification and characterization of a non-hemorrhagic metalloproteasefrom Agkistrodon halys brevicaudus venom, Biochim. Biophys. Acta 1243(1995) 94e100.

[33] A. Andrieux, G. Hudry-Clergeon, J.J. Ryckewaert, A. Chapel, M.H. Ginsberg,E.F. Plow, G. Marguerie, Amino acid sequences in fibrinogen mediating itsinteraction with its platelet receptor, GPIIbIIIa, J. Biol. Chem. 264 (1989)9258e9265.

[34] G.E. Jarvis, B.T. Atkinson, D.C. Snell, S.P. Watson, Distinct roles of GPVI andintegrin a2b1 in platelet shape change and aggregation induced by differentcollagens, Br. J. Pharmacol. 137 (2002) 107e117.

[35] W.J. Wang, C.H. Shih, T.F. Huang, Primary structure and antiplatelet mecha-nism of a snake venom metalloproteinase, acurhagin, from Agkistrodon acutusvenom, Biochimie 87 (2005) 1065e1077.

[36] I. Tanjoni, K. Evangelista, M.S. Della-Casa, D. Butera, G.S. Magalhães, C. Baldo,P.B. Clissa, I. Fernandes, J. Eble, A.M. Moura-da-Silva, Different regions of theclass P-III snake venom metalloproteinase jararhagin are involved in bindingto a2b1 integrin and collagen, Toxicon 55 (2010) 1093e1099.

[37] T. Escalante, N. Ortiz, A. Rucavado, E.F. Sanchez, M. Richardson, J.W. Fox,J.M. Gutiérrez, Role of collagens and perlecan in microvascular stability:exploring the mechanism of capillary vessel damage by snake venom met-alloproteinases, PLoS One 6 (2011) e28017.

[38] S. Marcussi, C.P. Bernardes, N.A. Santos-Filho, M.V. Mazzi, C.Z. Oliveira,L.F. Izidoro, A.L. Fuly, A.J. Magro, A.S. Braz, M.R. Fontes, J.R. Giglio, A.M. Soares,Molecular and functional characterization of a new non-hemorrhagic metal-loprotease from Bothrops jararacussu snake venom with antiplatelet activity,Peptides 28 (2007) 2328e2339.