Determination of molecular mass.

Molecular mass of daboxin P was determined by electrospray ionization mass spectrometry (ESI-MS) using LCQ fleet Ion Trap, Thermo Scientific (MA, USA). Ion spray voltage was set at 4.4 kV. For nebulization nitrogen gas was used. Solvent (50% MeCN containing 0.1% formic acid) was delivered into the system at a flow rate of 50 μl/min by Accela 600 pump. Spectra were recorded in positive mode with mass to charge (m/z) ranging from 800 to 2000. Analysis and deconvolution of the spectra was performed by Promas for Xcaliber.

Primary structure determination.

N-terminal sequencing: N-terminal sequencing of the native purified protein was performed by automated Edman degradation process using Protein sequencer PPSQ 31A Shimadzu Asia Pacific, (Singapore). Briefly, 89 μg (6545.56 pmol) of the purified protein (dissolved in 100 μl of milli q water) was dried on PVDF membrane and loaded onto the sequencer. Sequencing was performed for 30 cycles or till the chromatogram was readable.

Pyridylethylation: Prior to chemical cleavage, daboxin P was reduced and alkylated according to the method developed by Joseph and colleagues [29,30]. Briefly, 1 mg of daboxin P was dissolved in 920 μl of denaturing buffer (6 M guanidine hydrochloride and 0.1 M Tris-Cl pH 8.5) and treated with β-mercaptoethanol (30 μl). N₂ gas was purged over the reaction mixture to expel any dissolved oxygen and incubated at 37°C for 150 min. 50 μl of 4-vinyl pyridine (containing 100 ppm hydroquinone) was added and further incubated at 37°C for 150 min. The pyridylethylated sample was desalted by Rp-HPLC on Jupiter C₁₈ column (4.6 x 250 mm, 3 μm, 300 Å) with a linear gradient of 80% MeCN containing 0.1% TFA at a flow rate of 0.8 ml/min. Elution was monitored at 215 and 280 nm.

Chemical cleavage: The chemical cleavage of pyridylethylated protein by BNPS-skatole and hydroxylamine hydrochloride were carried out according to the protocol developed by Crimmins and co-workers and Milner and colleagues respectively with minor modifications [31,32]. For cleavage with BNPS-skatole, 900 μg of the pyridylethylated protein was dissolved in 0.1% of TFA containing 6 M guanidine hydrochloride (pH 5.0). To this 1.5 mg/ml of BNPS skatole (dissolved in 100% acetic acid) was added and incubated at 37°C for 24 h. The reaction was stopped by addition of equal volume of milli q water. Centrifugation was carried out at 12,000 rpm for 20 min and supernatant was collected.

For cleavage with hydroxylamine hydrochloride, 1 mg of pyridylethylated protein was dissolved in denaturing buffer (6 M guanidine hydrochloride, 0.1 M Tris-Cl, pH 9.0, 2 M hydroxylamine hydrochloride and 0.2 M K₂CO₃) and incubated at 45°C for 4 h. After the cleavage experiments, the samples were subjected to Jupiter C₁₈ column using a linear gradient of 5–50% of 80% MeCN containing 0.1% TFA. The mass of the fractions obtained were checked using ESI-MS (Accela LCQ fleet ion trap, Thermo scientific, U.S.A) for the cleaved and uncleaved peptide fragments. The cleaved fractions were lyophilized and reconstituted in 100 μl of milli q water and loaded onto PVDF membrane for Edman degradation using Automated Protein sequencer PPSQ 31A, Shimadzu.

Enzymatic cleavage/ Tandem mass spectrometry: Daboxin P (50 μg) was treated with 50 mM of ammonium bicarbonate, 1% Protease Max and 0.5 M dithiothreitol at 56°C for 20 min. For alkylation, 550 mM of iodoacetamide was added and incubated in dark for 15 min. Following this, trypsin (1 μg/μl in 50 mM acetic acid) in the presence of 1% Protease Max was added to the reaction and incubated at 37°C for 3 h. Reaction was stopped by 100%TFA. Centrifugation was carried out at 12,000 rpm for 10 min and the supernatant was subjected to Tandem mass spectrometry as described previously [11]. The MS/MS spectra obtained were analyzed using Proteome Discoverer 3.1 with Sequest program.

Sequence alignment and phylogenetic analysis.

PLA₂ enzymes having maximum sequence similarity with daboxin P were retrieved by blastp analysis from the NCBI database and aligned using multiple sequence alignment software DNAMAN 4.1.5.1 (Lynnon BioSoft). The phylogenetic relationship among the PLA₂ enzymes exhibiting anticoagulant activity were determined by the Mega 5.05 software using neighbor joining method with a bootstrap of 1000 replicates [33]. The evolutionary distance among the sequences was calculated with p-distance.

Secondary structure determination.

The secondary structure was analyzed by circular dichroism (CD) spectroscopy using Jasco Spectropolarimeter J-810 (Tokyo, Japan). Briefly, 200 μl of 0.4 mg/ml of daboxin P (dissolved in milli q water) was subjected to the spectropolarimeter. The spectrum of the native protein was recorded at far UV scan from 260–190 nm at a speed of 50 nm/min in quartz cuvette (0.1cm path length) at room temperature.

The effect of different pH (phosphate buffered saline at pH 3.0, 7.4 & 12) and temperature (from 25°C to 100°C) on structural conformation of daboxin P (0.4 mg/ml) was determined under the same experimental conditions as described above. Three scans were recorded for each experimental condition and subtracted from blank to obtain the final spectra of the protein.

PLA₂ activity.

The PLA₂ activity of daboxin P was assayed using the sPLA₂ assay kit (Cayman, MI, USA) according to the manufacturer’s protocol using diheptanoylthio-phosphatidylcholine as substrate. The specific activity was expressed in micromoles of phosphatidylcholine hydrolyzed per min per mg of enzyme. Bee venom PLA₂ enzyme (0.001 μg/μl) was considered as the positive control for the experiment. The results are mean ± standard deviation (SD) of three independent experiments.

Alkylation.

The alkylation of His48 of daboxin P was performed as described by Maduwage and co-workers [34]. Briefly, 150 μg of daboxin P was treated with 15 μl of p-Bromophenacyl Bromide (pBpB) (100 μg dissolved in 30 μl of 100% ethanol). Equal amount of the protein was treated with 15 μl of ethanol as the vehicle control. The reaction mixtures were incubated at room temperature for 20 h and subsequently dialyzed in 20 mM of Tris-Cl, pH 7.4 at 4°C for 3 times. The dialyzed sample was lyophilized in a pre-conditioned lyophilizer (112N-G, Hahntech, South Korea) at –80°C. The lyophilized sample was reconstituted in appropriate buffer to carry out the desired assays. The extent of alkylation was verified by analyzing it using ESI-MS.

Cytotoxicity study.

Cell culture: The HEK-293 and MCF-7 cell lines were grown in DMEM media enriched with 10% FBS and 1% antibiotic (Strep-Pen). The cells were grown on 96 well plate to a confluency of 70–80% at 37°C for 2–3 days in CO₂ (5%) incubator (Eppendorf, Hamburg, Germany). Cell viability was assessed by trypan blue stain and the cells were quantified on a haemocytometer.

The cytotoxic activity of the crude venom of Daboia r. russelii and daboxin P were determined by the MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) based colorimetric method [35]. Briefly, in two independent set of experiments, the HEK-293 and MCF-7 cells were plated on 96 well plate at a concentration of 1x10⁵ cells/well and incubated for 48 h at 37°C in CO₂ (5%) incubator. Subsequently, the cells were incubated with different concentration of crude Daboia r. russelii venom and daboxin P for 24 h at 37°C in CO₂ (5%) incubator. Cells treated with 0.9% NaCl was considered as the vehicle control. After incubation, the changes in morphological pattern (if any) of the treated cells were observed in an Inverted microscope (Axio Vert A1, Zeiss, Jena, Germany) at 10X magnification. 20 μl of MTT (5 mg/ml) was added and incubated at 37°C for 2 h. Following this, the formazan crystals formed by the viable cells were dissolved by 150 μl of MTT solution. The amount of formazan formed by the viable cells was quantified at 590 nm using MultiSkan Go spectrophotometer (Thermo Scientific, MA, USA). The percentage of cell viability was calculated by considering the cells without the venom treatment as 100% viable. The results are mean ± SD of three independent experiments.

In vitro-anticoagulant activities.

The anticoagulant experiments were carried out with platelet poor human plasma (PPP) (whole blood centrifuged at 3000 rpm for 20 min at 16°C) and clot formation was monitored using Tulip Coastat-1 coaguloanalyser (Alto Santa Cruz, Goa, India). For each assay, clotting time of plasma with Tris-Cl buffer (20 mM, pH 7.4) was taken as the normal clotting time. The results are mean ± SD of three independent experiments.

Recalcification time (RT): Different concentrations of daboxin P (0.001, 0.01, 0.1, 1 μM) were pre-incubated with 150 μl of plasma at 37°C for 2 min [36]. Formation of plasma clot was initiated by addition of 100 μl of 50 mM CaCl₂.

Activated partial thromboplastin time (APTT): APTT was determined using Liquecelin (APTT reagent) according to the manufacturer’s instructions. Briefly, different concentrations of the purified protein (0.01, 0.1, 1 μM) were pre-incubated with human plasma (50 μl) and APTT reagent (50 μl) for 3 min at 37°C. Plasma clot was formed by addition of 25 mM CaCl₂ (50 μl).

Prothrombin time (PT): PT was measured using Uniplastin (PT reagent) according to the instructions of manufacturer. In brief, different concentrations of daboxin P were pre-incubated with 50 μl of human plasma at 37°C for 2 min. PT reagent (50 μl) was added to initiate the clot formation.

Stypven time: Briefly, in a reaction volume of 300 μl different concentrations of daboxin P were pre-incubated with human plasma (75 μl) for 3 min at 37°C [36]. To this 75 μl of RVV-X (10 ng/ml) was added and incubated for 2 min at the same temperature. Clotting was initiated by addition of 25 mM of CaCl₂.

Thrombin time: To determine the thrombin time, different amounts of the protein (1, 3, 5 μg) were incubated with human plasma (50 μl) for 3 min at 37°C [36]. To this 50 μl of thrombin (10 u/ml) was added to initiate clot formation.

Fibrinogenolytic activity: Fibrinogenolytic activity was determined according to the method developed by Ouyang C and Teng CM [37]. In brief, different amounts of daboxin P were pre-incubated with 2 mg/ml fibrinogen (dissolved in 50 mM Tris-Cl, pH 7.5, 0.15 M NaCl) for 24 h at 37°C. For positive control, equal amount of fibrinogen was incubated with thrombin (3 μl of 10 units/ml) while for negative control fibrinogen with 50 mM Tris-Cl, pH 7.5 was incubated under the same experimental conditions. The formation of fibrinogen degradation products were analyzed on 12.5% glycine SDS-PAGE stained with Coomassie Brilliant Blue R-250.

In-vivo anticoagulant activity.

The in-vivo anticoagulant activity of daboxin P was determined by the FeCl₃-induced carotid artery thrombosis model with minor modifications [38–40]. Briefly, C57BL/6 male mice (9–11 weeks old, 24–28 g) were anesthetized with ketamine (75 mg/kg) and medetomidine (1 mg/kg) by intraperitoneal injection (i.p.). Daboxin P (10 mg/kg) was injected into the mice (n = 6) via its tail vein. Following this, the right carotid artery was exposed using blunt dissection and a doppler flow probe (Model MA0.5VB, Transonic System Inc., Ithaca, NY, USA) connected to a perivascular flow module (TS420, Transonic System Inc., Ithaca, NY, USA) was then attached to the carotid artery to monitor blood flow. A 2x2 mm piece of filter paper (soaked in 10% FeCl₃ solution) was placed on the surface of the carotid artery for 3 min to initiate the thrombus formation. Saline treated mice (n = 5) were considered as the negative control for the experiment. Time-to-occlusion (TTO) is defined as the time taken for the blood flow to reach zero after the application of FeCl₃. Maximum measurement time was considered for 60 min after the application of FeCl₃. TTO was recorded as 60 min, if no occlusion occurred by this time. The animal experiments were approved by Institutional Animal Care and Use Committee, National University of Singapore

Serine protease specificity.

The effect of daboxin P on the amidolytic activity of serine proteases were determined using specific chromogenic substrates. The activated serine proteases involved in the extrinsic and intrinsic tenase complex namely, FXIIa (60 nM), FXIa (0.125 nM), FIXa (333 nM), FXa (0.43 nM), FVIIa (10 nM), sTF (30 nM) were reconstituted in the activation buffer containing 50 mM Tris-Cl pH 7.4, 1 mg/ml of BSA, 140 mM NaCl and 5 mM CaCl₂ [41]. The respective chromogenic substrates, S-2302 (1 mM), S-2366 (1 mM), spectrozyme (0.4 mM), S-2765 (0.65 mM) and S-2288 (500 μM) were reconstituted in milli q water to the desired working concentration. Briefly, to a reaction volume of 200 μl, each of the serine proteases was pre-incubated with different concentration of daboxin P (0.01, 0.1 and 1 μM) for 5 min. Subsequently, respective chromogenic substrates were added to trigger the amidolytic reaction. For each assay, the rate of p-nitroaniline formation upon hydrolysis of the substrate by the specific enzyme was quantified at 405 nm using UV-Vis MultiSkan GO spectrophotometer. The hydrolysis of substrate in the absence of daboxin P was considered as 100%.

Reconstitution of the tenase complexes.

Extrinsic tenase complex (ETC) with and without phospholipids: The extrinsic tenase complex was reconstituted under in-vitro conditions in the presence of FVIIa (5 nM) and Innovin (10X) (tissue factor, thromboplastin and Ca⁺² ions) with 67 μM of phospholipid vesicles (DOPC:DOPS 7:3) in the activation buffer. The reaction mixture was incubated at 37°C for 15 min [20]. Similarly, the complex was formed without phospholipids with 20 nM of FVIIa and soluble tissue factor, sTF (60 nM) and incubated for 15 min at 37°C.To both the experiments, different concentrations of daboxin P (0.01, 0.1, 1 and 3 μM) were added and incubated for 15 min. Following this, the reactions were incubated with FX (30 nM) for 15 min. Both the experiments were stopped by addition of quenching buffer (50 mM Tris-Cl pH 7.4, 1 mg/ml of BSA, 140 mM NaCl and 50 mM EDTA). The amount of FXa generated was determined by addition of 500 μM of S-2222 and the rate of hydrolysis was quantified at 405 nm using UV-Vis MultiSkan GO spectrophotometer.

Intrinsic tenase complex with and without phospholipids (ITC): Reconstitution of the intrinsic tenase complex under in-vitro conditions was carried out in the presence of phospholipids (67 μM) in activation buffer with FVIIIa, FIXa, FX and Ca⁺² ions at 37°C [42,43]. Briefly, 3 nM of FVIII was incubated with 500 pM of thrombin for 10 min to activate FVIII. Thrombin was then deactivated by 10 nM of variegin, a thombin inhibitor from Amblyomma variegatum [44]. After incubation of the reaction mixture with 1 nM of FIXa for 10 min different concentrations of daboxin P (0.01, 0.1, 1 and 3 μM) were added and incubated for 15 min. Following this, 25 nM of FX was added and incubated for 15 min. Similarly, the complex was formed devoid of the phospholipid blend [42] with FVIII (40 nM) and thrombin (4 nM) and incubated for 10 min. Thrombin was deactivated with 40 nM of variegin. To this 10 nM of FIXa was added and incubated for 10 min. Then, different concentrations of daboxin P (0.01, 0.1, 1 and 3 μM) were incubated for 15 min. Finally, 1 μM of FX was added and incubated for 15 min. The reactions in both the experiments were terminated by addition of the quenching buffer. The amount of FXa generated was quantified by the hydrolysis of the chromogenic substrate S-2222 (500 μM) at 405 nm using UV-Vis MultiSkan GO spectrophotometer.

For determining the inhibitory concentration (IC50) of daboxin P on extrinsic and intrinsic tenase complex, different concentration of daboxin P were pre-incubated with each of the reconstituted complexes described above in the presence of phospholipids in the activation buffer.

Prothrombinase complex: Prothrombinase complex was reconstituted in the presence of phospholipid (67 μM) with FXa, FVa, and Ca⁺² ions under in-vitro conditions at 37°C [19]. Briefly, 10 pM of FXa was pre-incubated with 1 nM of FVa for 15 min. To this, different concentrations of daboxin P (0.01, 0.1, 1 and 3 μM) were added and incubated for 15 min. Thereafter, prothrombin (12.5 nM) was added and the reaction mixture was incubated for another 15 min. In another set of experiment, FXa (10 pM) was pre-incubated with different concentrations of daboxin P followed by incubation with FVa (1 nM) and prothrombin (12.5 nM) for 15 min each. For both the experiments, the reactions were stopped with quenching buffer. The amount of thrombin generated by the complex was determined by the rate of hydrolysis of the chromogenic substrate S-2238 (250 μM) at 405 nm using UV-Vis MultiSkan GO spectrophotometer.

Fluorescence emission spectroscopy.

Interaction of daboxin P with the coagulation factor FX and FXa were analyzed using fluorescence emission spectrophotometer (LS 55, Perkin Elmer, Palo Alto, CA). In brief, 1 μM of daboxin P (dissolved in 20 mM Tris-Cl, pH 7.4) was incubated with either 0.05 μM of FX or 0.1 μM of FXa (reconstituted in the same buffer) for different time intervals (10 min & 20 min) at room temperature in the presence and absence of 5 mM CaCl₂. The change in the fluorescence emission spectra of the individual components and the mixtures were measured over a wavelength of 200 to 500 nm with an excitation wavelength of 280 nm at room temperature using a quartz cuvette (1 cm path length).

Affinity chromatography.

The association of daboxin P with FX and FXa were evaluated by affinity column chromatography using CNBr activated sepharose® 4B (Sigma, Aldrich). Briefly, 40 μg of daboxin P was covalently linked to 0.1 g of CNBr activated matrix (swollen overnight in ice cold 1 mM HCl) in the presence of coupling buffer containing 0.1 M NaHCO₃ and 0.5 M NaCl, pH 8.3. Subsequently, the unbound protein molecules of daboxin P were washed away with the same buffer. To avoid non-specific binding, the immobilized matrix was treated with 0.2 M glycine for 2 h with gentle shaking at room temperature. The blocking solution was removed by an alternative wash with acidic (0.1 M acetate buffer, pH 4.0) and basic buffer (coupling buffer) for 5 times. FX or FXa (8 μg) were added to the daboxin P immobilized matrix and incubated for 1 h at room temperature with mild shaking. The unbound FX or FXa molecules were washed away with the same buffer. Elution of FX or FXa was carried out with 20 mM HCl containing increasing molar concentrations of NaCl (0.5, 1, 1.5 and 3 M). The flow through and eluents were analyzed on 12.5% glycine SDS-PAGE under non-reducing condition and visualized after staining with Pierce^TM silver staining kit.

Modelling and molecular docking studies.

Three dimensional (3D) molecular modelling: For the in-silico 3D modelling of daboxin P, online server I-TASSER was employed [45–47]. This server predicts the secondary conformation of a given protein by profile-profile alignment (PPA) threading techniques. The model with maximum C-score and TM score was selected for generation of ribbon model using DS ViewerPro 5.0 (Accelrys Inc.) software. Superimposition of the predicted model with the crystal structure of ammodytoxin A, AtxA (Pdb 3G8G), was carried out using DS ViewerPro 5.0 software.

Molecular docking: The equilibrated model of daboxin P was docked with FXa (PDB: 2BOH) using the online web server, PatchDock [48,49]. This server uses the geometry-based docking algorithm to find optimum candidate solutions and the RMSD (root mean square deviation) clustering to remove redundant solutions [49]. Each solution was given score based on the geometric fit as well as atomic desolvation energy [50]. In the present analysis, a default RMSD value of 4 Å was considered for clustering solutions. The ribbon model of the docked complex of daboxin P and FXa was generated using DS ViewerPro 5.0 (Accelrys Inc.).

Identification of interface residues and hot spot residues: The daboxin P-FXa docked model with maximum surface contact area and minimum free energy generated by PatchDock server was selected for evaluating the interface residues using PDBsum server [51,52]. The interface residues of a protein are those residues whose contact distances from the interacting protein partner are less than 6 Å [53].

Contact map analysis: The interaction between the chains of FXa and daboxin P were also analyzed using CMA (Contact Map Analysis), one of the servers of the online software SPACE (Structure Prediction and Analysis based on Complementarity with Environment) [54]. The server analyses the contact area between the residues of two protein chains or within a single chain of a given PDB file based on shape and chemical complementarity. It displays residue to residue contacts for a pair of amino acid residues involved in the interaction in the form of contact map. Contact maps provide more concrete demonstration of protein structure than its 3D atomic coordinates. For investigating the residue-residue contacts between daboxin P and FXa, a contact area threshold above 8Ų has been considered for the present analysis.

Sequencing.

The first 30 residues were determined by N-terminal sequencing (Fig 3A). The cleavage of pyridylethylated daboxin P with BNPS-skatole yielded two peptide fragments of molecular mass 11.7 kDa and 3.6 kDa (data not shown) while cleavage with hydroxylamine hydrochloride resulted into two peptide fragments of molecular mass 6.6 kDa and 8.6 kDa (data not shown). Sequencing of these peptides by Automated Edman degradation revealed the rest of the amino acid residues (Fig 3A). On the other hand from tandem mass spectrometry, eleven peptide fragments of daboxin P were obtained (Table A in S1 File). These peptide fragments obtained by chemical and enzymatic cleavage were assembled and overlapped to decipher the complete sequence of daboxin P (Fig 3A). The homology of the protein sequence was analyzed by multiple sequence alignment with PLA₂ enzymes using online blastp algorithm as represented in Fig 3B.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Primary structure of daboxin P.

(A): Amino acid sequence was deciphered by Edman degradation sequencing and ESI-LC MS/MS. The peptide sequences obtained from N-terminal sequencing and chemical cleavage by BNPS-skatole and hydroxylamine hydrochloride are shown with two headed solid arrows whereas peptide sequences obtained after ESI-LC MS/MS of the tryptic digested fragments are indicated with two headed doted arrows. (B): Multiple sequence alignment of daboxin P with the PLA₂ enzymes from different subspecies of Daboia russelii (24638087: D. r. russelii, 408407675: D. r. siamensis, 31615955: D. r. pulchella, 49259309: D. r. russelii, 31615954: D r.pulchella, 109157490: D. r. pulchella, 48425253: D. r. pulchella, 298351762: D. r. russelii, 81174981: D. r. russelii). The conserved cys residues are highlighted in grey and the amino acid substitutions in daboxin P are underlined. * indicates the His residue at the active site. The predicted anticoagulant region is highlighted with a solid black line.

https://doi.org/10.1371/journal.pone.0153770.g003

Sequence analysis.

The calculated molecular mass of daboxin P was found to be 13596.70 Da which is in agreement with the observed mass of 13597.62 ± 1.28 Da from ESI-MS. It has 121 amino acid with 14 cysteine residues which should correspond to 7 disulphide bonds, a characteristic feature of group IIA viper PLA₂ enzymes [55,56]. It has a theoretical pI of 8.5, calculated by online software ExPASy ProtParam [57]. At the active site, histidine residue was observed followed by an aspartate residue, a common trait for Ca⁺² dependent catalytic PLA₂ enzymes [58]. At the predicted anticoagulant region (53NLPDCNNKSKRYRYKK68) positively charged amino acid residues (Lys and Arg) were observed which propose it to be a strong anticoagulant PLA₂ enzyme.

Homology search by blastp shows its similarity to VRV-PL-VIIIa (96%) isolated from Daboia r. russelii at position 59^th (Asn), 62^nd (Lys), 65^th (Arg), 68^th (lys) and 120^th (Val) (Fig 3B) [59].

Phylogenetic relationship.

Daboxin P showed close phylogenetic relation with the reported anticoagulant PLA₂ enzymes which are basic, like vurtoxin which is a strong anticoagulant enzyme reported from Vipera renardi and FXa binding enzymes like DPLA2 from Daboia russelii pulchella, AtxA and AtxC from V. a. ammodytes with a bootstrap value of 99 [59–61] (Figure A in S1 File).

Secondary structure determination.

The far UV CD spectrum has shown negative minima at 222 nm and 208 nm and positive maxima at 190 nm, typical for α-helical pattern of PLA₂ enzymes (Fig 4A). The percentage of secondary structure was determined by online software K2D3 which calculated 42.73% of α-helix and 12.36% of β-sheet in daboxin P [62,63].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Far-UV circular dichroism (CD) spectra of daboxin P (0.4 mg/ml).

(A): in milli q water at 25°C, (B): at different pH (3.0, 7.4 & 12) at 25°C. (C) Melting curve of daboxin P (dissolved in milli q water) at 222 nm considering temperature as a function. The curve was plotted using sigmoidal curve fit and Tm value was determined by Boltzman equation using Origin (OriginLab).

https://doi.org/10.1371/journal.pone.0153770.g004

The CD spectrum at acidic pH (3.0) and neutral pH (7.4) showed stable structural conformation. However, the structure was distorted completely at alkaline pH (12.0) with residual α-helix of 5.51% & β-sheet of 23.75% (Fig 4B). On the other hand, its secondary structure showed a Tm (melting temperature) value of 71.59 ± 0.4°C when scanned over a range of temperature (Fig 4C).

Catalytic activity.

Daboxin P exhibited catalytic activity on diheptanoylthio-phosphatidylcholine (PC) in a dose dependent manner (Fig 5A). The Michaelis Menten plot with different substrate concentration showed a hyperbolic curve for enzyme activity and the Km (6.6 mM) and Vmax (1.14 mmol/(min*mg)) were determined using Lineweaver Burk-plot (Fig 5B & 5C). Alkylation with pBpB has led to the complete loss of its enzymatic activity signifying the crucial role of histidine residue in catalysis (Fig 5A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Phospholipase A₂ activity of daboxin P.

(A) Progress curve of diheptanoyl thio-PC cleavage by daboxin P, bee venom PLA₂ enzyme and histidine modified daboxin P* at 414 nm. (B) Michaelis-Menten’s curve for sPLA₂ assay. (C) The Lineweaver-Burk plot of daboxin P for determination of Km and Vmax.

https://doi.org/10.1371/journal.pone.0153770.g005

Cytotoxicity.

The crude venom of Daboia r. russelii has shown prominent cytotoxic effect on HEK-293 and MCF-7 cells in a dose dependent manner (Fig 6). At a concentration of 2.8 and 3 μg/ml of crude venom, only 13.09% of HEK-293and 14.97% of MCF-7 cells respectively (Fig 6B, 6E, 6G and 6H) remained viable while the same concentrations of daboxin P did not show any cytotoxic effect, suggesting that the cytotoxic effect of the crude venom is not due to this protein (Fig 6C and 6F–6H).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Cytotoxic effect of crude Daboia r. russelii venom and daboxin P.

Microscopic images were photographed at 10X magnification under Inverted microscope (Axio Vert A1., Zeiss) after treatment with venom samples for 24 h (A): HEK-293 cells treated with 0.9% NaCl were considered as negative control (B): HEK-293 cells treated with crude Daboia r. russelii venom (5 μg/ml) (C): HEK-293 cells treated with daboxin P (5 μg/ml) (D): MCF-7 cells treated with 0.9% NaCl were considered as negative control (E): MCF-7 cells treated with crude Daboia r. russelii venom (5 μg/ml) (F): MCF-7 cells treated with daboxin P (5 μg/ml). Percentage cell viability (G): HEK-293; (H): MCF-7 after treatment with crude venom and daboxin P using MTT based colorimetric assay. Percentage cell viability was calculated by considering the cells without venom treatment as 100% viable. The results are mean ± SD of three independent experiments.

https://doi.org/10.1371/journal.pone.0153770.g006

In-vitro anticoagulant activity.

Daboxin P exhibited anticoagulant effect on the various coagulation assays in a dose dependent manner (Fig 7). At a concentration of 0.01 μM, it prolonged the recalcification time beyond 600 s indicating it to be a strong anticoagulant enzyme (Fig 7A). 1 μM of the protein delayed activated partial thromboplastin time up to 130 s but did not exhibit prominent effect on prothrombin time (Fig 7B & 7C). Daboxin P exhibited anticoagulant effect on stypven time in a dose dependent manner displaying its inhibitory effect on FX (Fig 7D). Nevertheless, it did not show any inhibitory effect on thrombin time and was devoid of fibrinogenolytic activity (Figure B i & ii in S1 File). This suggests that daboxin P might target a component/coagulation factor upstream of the common pathway.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Anticoagulant activities of daboxin P on platelet poor human plasma (PPP).

(A): Recalcification time, different concentrations of daboxin P (0.001, 0.01, 0.1) were pre-incubated with 150 μl of plasma at 37°C for 2 min. 100 μl of 50 mM CaCl₂ was added to initiate clot formation. (B): Activated partial thromboplastin time, daboxin P (0.01, 0.1 & 1 μM) was pre-incubated with 50 μl of plasma and 50 μl APTT reagent (Liquecelin) for 3 min at 37°C. 50 μl of 25 mM CaCl₂ was added to form clot. (C): Prothrombin time, different concentrations of daboxin P (0.01, 0.1 & 1 μM) were pre-incubated with 50 μl of plasma at 37°C for 2 min. 50 μl of PT reagent (Uniplastin) was added to initiate the clot formation. (D): Stypven time, daboxin P (0.01, 0.1 & 1 μM) was pre-incubated with 75 μl plasma for 3 min at 37°C. 75 μl RVV-X (10 ng/ml) was added and incubated for 2 min. 25 mM of CaCl₂ was added to initiate clot formation. For all the experiments, the clot formation was monitored using Tulip Coastat-1 coagulo analyser and the time taken for clot formation in the presence of Tris-Cl buffer (20 mM, pH 7.4) was considered as normal clotting time (NCT). The results are mean ± SD of three independent experiments.

https://doi.org/10.1371/journal.pone.0153770.g007

In-vivo anticoagulant activity.

Daboxin P exhibited anticoagulant effect in the carotid artery of mouse treated with FeCl₃. Mice administered with saline had a time-to-occlusion within 10 min. However, daboxin P administration resulted in ~4-fold increase (35.18 ± 21.58 min) in the time-to-occlusion in comparison to the saline treated mice (8.29 ± 2.61 min) (Fig 8). This result was in accordance with the in-vitro anticoagulant assays on human plasma and highlighted the anticoagulant property of daboxin P under in-vivo condition.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. Effect of daboxin P on the time to occlusion (TTO) in FeCl₃ induced carotid artery thrombosis in mice.

C57BL/6 male mice anesthetized with ketamine (75 mg/kg) and medetomidine (1 mg/kg) (i.p) were injected (i.p.) with daboxin P (10 mg/kg) in tail vein. Saline treated mice were considered as negative control. Each data-point represents the time-to-occlusion (TTO) of a single mouse. Maximum experimental time was considered for 60 min after FeCl₃ induction.

https://doi.org/10.1371/journal.pone.0153770.g008

Serine protease specificity.

Screening for inhibitory effect of daboxin P on the amidolytic activity of various serine proteases involved in the extrinsic (FVIIa), intrinsic (FXIIa, FXIa, FIXa) and common (FXa) pathway revealed that it did not inhibit any of these serine proteases when assayed using respective synthetic chromogenic substrates (Fig 9A). However, it exerted inhibitory effect on both the extrinsic (ETC) (IC50 = 0.43 nM) and intrinsic (ITC) (IC50 = 39.20 nM) tenase complexes causing hindrance in the activation of FX to FXa (Fig 10). The residual activity of ETC and ITC was observed to be 9.05% ± 1.9 and 2.8% ± 1.6 respectively upon treatment with 3 μM of daboxin P (Fig 9B & 9C). Further, when the tenase complexes were reconstituted without phospholipid, the residual activity of ETC and ITC was found to be 25% ± 2.7 and 20.05% ± 2.3 respectively. Moreover, when the tenase complexes were treated with alkylated daboxin P (3 μM), only 23% ± 1.8 (ETC) and 19.03% ± 1.8 (ITC) of residual activity remained which is comparable to the residual activity of both the complexes reconstituted without phospholipids (Fig 9B & 9C). Interestingly, the thrombin formation by the prothrombinase complex was unaffected when daboxin P was added after the formation of the FXa-FVa complex. However, only 11.28% of residual activity remained when it was pre-incubated with FXa followed by addition of FVa for the formation of prothrombinase complex (Fig 9D).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 9. Percentage residual amidolytic activity of various serine proteases and complexes pre-incubated with daboxin P.

(A): Residual activity of FXIIa, FXIa, FXa, FIXa and FVIIa; (B): Activity of extrinsic tenase complex (ETC) (C): intrinsic tenase complex (ITC), in the presence or absence of phospholipid and alkylated daboxin P (indicated by *). (D): Residual activity of prothrombinase complex. Daboxin P was either pre-incubated with FXa followed by addition of FVa (pre-complex) or after reconstitution of FXa-FVa complex (post-complex). The rate of hydrolysis of respective chromogenic substrates for all the assays was measured at 405 nm using Multiskan Go spectrophotometer. Activity of the serine protease/complex without daboxin P was considered as 100%. The results are mean ± SD of three independent experiments.

https://doi.org/10.1371/journal.pone.0153770.g009

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 10. Inhibition curve (IC50) of daboxin P for the extrinsic and intrinsic tenase complex.

Different concentrations of daboxin P (1x10^-6, 1x10^-5, 1x10^-4, 1x10^-3, 1x10^-2, 1x10^-1, 1, 3x10^-6, 3x10^-5, 3x10^-4, 3x10^-3, 3x10^-2, 3x10^-1 and 3 μM) were pre-incubated with reconstituted tenase complexes as described above. The IC₅₀ was calculated by fitting the points by non-linear curve fit using Origin (OriginLab, Northampton, MA).

https://doi.org/10.1371/journal.pone.0153770.g010

Fluorescence emission spectroscopy.

The fluorescence emission spectrum of FX and FXa was quenched by daboxin P with increasing incubation time (Fig 11A & 11B). The presence or absence of Ca⁺² ions did not show any effect on the quenching spectra for both FX and FXa when treated with daboxin P (data not shown).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 11. Interaction of daboxin P with FX and FXa.

(A): Fluorescence emission spectra of daboxin P, FX, and the complex (daboxin P + FX). (B) Fluorescence emission spectra of daboxin P, FXa, and the complex (daboxin P + FXa). 1 μM of daboxin P (DP) was pre-incubated with either 0.05 μM of FX or 0.1 μM of FXa for different time intervals (10 min & 20 min) at room temperature. The emission spectra of the individual proteins and the complexes were measured from 200 to 500 nm with an excitation wavelength of 280 nm using quartz cuvette (1 cm path length). (C): Electrophoretic profile of the flow through and elute obtained after affinity column chromatography. Lane 1: PageRuler^TM Plus pre-stained protein ladder (250–10 kDa), Lane 2: flow through (FX) Lane 3: FX after elution with 1.0 M NaCl, Lane 4: control (FX); Lane 5: blank; Lane 6: PageRuler^TM Plus pre-stained protein ladder (250–10 kDa), Lane 7: flow through (FXa); Lane 8: FXa after elution with 1.0 M NaCl, Lane 9: control (FXa).

https://doi.org/10.1371/journal.pone.0153770.g011

Affinity column chromatography.

The electrophoretic profile of the flow through for both FX and FXa did not show any protein bands suggesting the binding of these proteins to the daboxin P immobilized CNBr matrix. Nevertheless, the eluent at 1.0 M NaCl showed prominent protein bands for both the serine proteases suggesting the disturbance of the protein-protein interaction at high salt concentration (Fig 11C).

Molecular modelling and protein-protein docking.

The predicted 3D model of daboxin P has shown structural similarity to the crystal structure of AtxA (PDB: 3G8G) with a RMSD value of 0.9 Å which validates the predicted model (Figure C i in S1 File). The conserved critical regions and amino acid residues in daboxin P are shown in Figure C ii in S1 File.

The best protein-protein docked model of daboxin P and FXa chosen in terms of minimum free energy and maximum surface contact area, was obtained from PatchDock web server (Fig 12). Daboxin P interacts with the heavy chain (B-Chain) of FXa with a geometric shape complementarity score of 12234. Geometric scoring refers to good molecular shape complementarity between the docked chains due to optimal fit with wide interface areas and lesser steric clashes [49]. The approximate interface area of the complex was found to be 1837 Ų with atomic contact energy (ACE) of 472.62 kcal/mol.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 12. 3D ribbon model of the docked complex of FXa and daboxin P.

The interface surface residues involved in the interaction were predicted by PDBsum. The Ca⁺² binding loop (Trp30, Gly31, Gly32); helix C (Asp48, Tyr51, Gly52); anticoagulant region (Asn58) and C-terminal region (Phe113) of daboxin P interact with the heavy chain of FXa (Thr132, Arg165, Lys169, Asn166, Leu170, Tyr225 and Arg125) are represented in scaled ball and stick. Inset: Diagram illustrating the interaction of the seven residues of chain B (heavy chain of FXa) with eight residues of chain C (daboxin P) as predicted by PDBsum server. Orange line denotes non-bonded contacts and blue line denote hydrogen bond.

https://doi.org/10.1371/journal.pone.0153770.g012

The interface and possible interacting residues across the interface of the complex were predicted by PDBsum. The total number of interface residues in protein-protein complex was found to be 31 and the interface area for each chain involved in the interaction was observed to be more than ~500 Ų. The docked complex was stabilized by molecular interactions like hydrogen bonding and non- bonded contacts. The interface amino acid residues of FXa (chain B) and daboxin P (chain C) involved in the interaction are shown in Fig 12 inset and the interface plot statistics is summarized in Table B in S1 File.

The docked complex shows interaction of daboxin P with the heavy chain (Chain B) of FXa at eight critical regions some of which are reported to be crucial for FVa binding [64]. The daboxin P residues Trp30, Gly31, Gly32 of the Ca⁺² binding loop; Asp48, Tyr51, Gly52 of helix C; Asn58 of the anticoagulant region and Phe113 of C-terminal region interact with residues Thr132, Arg165, Lys169, Asn166, Leu170, Tyr225 and Arg125 of heavy chain of FXa (Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Critical residues of daboxin P and FXa involved in interaction based on PDBsum analysis.

https://doi.org/10.1371/journal.pone.0153770.t001

Contact map analysis shows fourteen residues of daboxin P share a contact area greater than 8 Ų with sixteen residues of the heavy chain of FXa while only four residues of daboxin P are found to share a contact area greater than the threshold limit (8 Ų) with three residues of the light chain of FXa. This suggests the interaction of daboxin P with the light and heavy chain of FXa (Fig 13) (Table C in S1 File).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 13. Contact map of daboxin P-FXa complex generated using online server Contact Map Analysis.

(A): residue to residue contact of light chain of FXa and daboxin P (B): residue to residue contact of heavy chain of FXa and daboxin P. The residue to residue contact area of the interacting amino acid residues for the chains has been considered above 8 Ų for the design of the contact map.

https://doi.org/10.1371/journal.pone.0153770.g013