Purification and characterization of the first γ-phospholipase inhibitor (γPLI) from Bothrops jararaca snake serum

Phospholipases A2 (PLA2) are enzymes acting on the cell membrane phospholipids resulting in fatty acids and lysophospholipids and deconstructing the cell membrane. This protein is commonly found in snake venoms, causing tissue inflammation in the affected area. Evidence indicates that snakes have natural resistance to their own venom due to protective properties in plasma, that inhibit the action of proteins present in their venom. Given that, this study aimed to purify and characterize a γPLI from Bothrops jararaca serum, named γBjPLI. PLA2 inhibitor was isolated using two chromatographic steps: an ion exchange column (DEAE), followed by an affinity column (crotoxin coupled to a CNBr-activated Sepharose resin). The purity and biochemical characterization of the isolated protein were analyzed by RP-HPLC, SEC, SDS-PAGE, circular dichroism and mass spectrometry. The ability to inhibit PLA2 was determined by enzymatic activity, neutralization of paw edema and myonecrosis. The protein purity was confirmed by RP-HPLC and SEC, whilst an apparent molecular mass of 25 kDa and 20 kDa was obtained by SDS-PAGE, under reducing and non-reducing conditions, respectively. According to mass spectrometry analysis, this protein showed 72% and 68% of coverage when aligned to amino acid sequences of two proteins already described as PLIs. Thus, the inhibitory activity of enzymatic, edema and myonecrotic activities by γBjPLI suggests a role of this inhibitor for protection of these snakes against self-envenomation.


Introduction
Snakebite envenoming is classified as a neglected tropical disease by the World Health Organization [1]. Bothrops genus (Viperidae family) is responsible for~86% of snake accidents in PLOS

Ethics statement
The Laboratory of Herpetology of Butantan Institute, São Paulo (Brazil), supplied blood of B. jararaca and venom of Crotalus durissus terrificus (C. d. terrificus). The Animal House of Butantan Institute, São Paulo (Brazil) supplied specimens of Swiss mice. The ethical Committee for the Use of Animals of Butantan Institute approved these experimental protocols (1374/ 15 CEUAIB). After the experiments, the animals (female Swiss mice) were sacrificed with CO 2 . This proposal is in accordance with standards outlined by Brazilian laws for use of experimental animals, and with ethical principles adopted by the Brazilian College of Animal Experimentation (COBEA).

B. jararaca blood collection and serum separation
B. jararaca blood (10 specimens) was collected by caudal venipuncture and maintained at 4˚C overnight. Afterwards, the serum was obtained by centrifugation at 1200 × g for 15 min at 4˚C, and stored at -20˚C.

C. d. terrificus venom
Pooled lyophilized venoms of C. d. terrificus were supplied by the Laboratory of Herpetology at Butantan Institute, São Paulo (Brazil).

Purification of the crotoxin and PLA 2
Crotoxin was purified by fractionation of C. d. terrificus venom by size-exclusion chromatography, using a Sephacryl S200 HR column. Lyophilized crude venom pool of C. d. terrificus snakes (449 mg), was dissolved in 2.5 mL of 50 mM Tris, 0.1 M NaCl buffer (pH 7.4), centrifuged at 4500 × g for 15 min at 4˚C and filtered by 0.45 microfilter. The supernatant was applied on the column, previously equilibrated with 50 mM Tris, 0.1 M NaCl, pH 7.4. The PLA 2 was separated from crotapotin by Reverse Phase-High Performance Chromatography (RP-HPLC) using a Supelco C5 column (0.10 cm × 25 cm), as previously described by Toyama et al. [46].

Purification of the γBjPLI
γBjPLI was isolated from B. jararaca serum by a combination of two chromatographic steps: an anion exchange and afterwards an affinity chromatography. The purity of the isolated molecule was evaluated by RP-HPLC on C18 column. Anionic exchange chromatography. Seven milliliters of B. jararaca serum were diluted in 7 mL of 25 mM Tris buffer, pH 7.5 (buffer A) and injected onto a DEAE Fast Flow column (1.6 x 2.5 cm) (GE Healthcare), previously equilibrated with 95% of buffer A and 5% of buffer B (25 mM Tris, 1 M NaCl, pH 7.5). The elution was performed by two steps, initially using buffer A containing 100 mM NaCl, followed by a linear gradient of 100 mM to 500 mM NaCl. The last eluted peak (named D2) was dialyzed three times against PBS (using a membrane with a 12 kDa cut-off molecular weigth) during 16 h at 4˚C.
Affinity chromatography. Crotoxin (20 mg), isolated from C. d. terrificus venom as described in item 2.4, was coupled to 2 g of CNBr-activated Sepharose as described by the manufacturer (GE Healthcare). At the end of the coupling process, a wash sequence is made with acidic (pH 4) and basic (pH 8.3) buffers to remove molecules that would not be bound to resin correctly. It was then settled in a glass column and equilibrated with PBS. A control chromatography without serum sample was performed, in order to verify if crotoxin molecules γ-phospholipase inhibitor (γPLI) from Bothrops jararaca snake serum were being detached from the resin. For the second purification step of γBjPLI, D2 peak (from anionic exchange) was applied on the affinity column (crotoxin + CNBr-activated Sepharose). The column was washed with PBS again, and the inhibitor was eluted with 1 M glycine, pH 2. Fractions of 2 mL were collected, and the pH was immediately neutralized with 1 M Tris buffer, pH 8.8.

RP-HPLC
γBjPLI purity was analyzed using RP-HPLC on C18 column (Discovery BIO Wide Pore C18 HPLC Column, 25 cm × 4.6 mm). In this chromatography, approximately 1 mg of γBjPLI was dissolved in solution A (TFA 0.1%) that was also used for RP-HPLC equilibration for 15 min, before injection of samples. The elution of all samples was done using a linear gradient (0-100%) of solution B (66% ACN in solution A) and chromatographic run was conducted at 280 nm, at a constant flow rate of 1 mL/min for 40 minutes.

Size exclusion chromatography protein analysis (SEC)
In order to perform the protein analysis by SEC, samples of isolated PLA 2 , purified γBjPLI and PLA 2 incubated with γBjPLI (PLA 2 + γBjPLI) were prepared by dissolution of each protein in 0.05 M Tris-HCl buffer, pH 7.6 at a final protein concentration of 1 mg/mL. The chromatographic system composed by analytical SEC column separation was performed on silicabased column BioSep SEC S-2000 from Phenomenex (5 μm, 300 × 7.8 mm), maintained at 30˚C, coupled to analytical chromatography system Jasco. The chromatographic column was equilibrated with 0.05 M Tris-HCl buffer, pH 7.6 for 20 min before injection of 10 μL of each sample. The chromatographic run was monitored at 280 nm, at a flow rate of 1 mL/min and all fractions were recovered, lyophilized and stored for further investigations. For size comparisons, a Gel Filtration Standart (BioRad) was used, at the ranges from 670 kDa to 1.35 kDa.

Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis (SDS-PAGE)
SDS-PAGE using 15% polyacrylamide gels was performed according to Laemmli [47], in the presence or absence of reducing agent (β-mercaptoethanol). Gels were electrophoresed at constant amperage of 20 mA and the samples were stained with Coomassie brilliant blue G-250.

Mass spectrometry
Protein bands were excised from SDS-PAGE and subjected to reduction (10 mM dithiothreitol), alkylation (50 mM iodoacetamide), and overnight in-gel digestion with sequencing grade trypsin (Sigma), in 50 mM ammonium bicarbonate at 37˚C. Tryptic peptides were extracted with 1% acid formic and analyzed by nanoAcquity UPLC, coupled to a Synapt G2 HDMS mass spectrometer (Waters). The ion source was made by ESI (nano-spray), fragmentation by CID and CAD (y and b ions), MS scan by Quadrupole and MS/MS scan by Time of Flight (TOF). Fragment mass error tolerance was of 0.1 Da. A peak list was generated, and used to search against the "Uniprot_serpentes_unr" database. The alignment of the amino acid sequence obtained by mass spectrometry analysis was done using the program Clustal Omega 1.2.4 [48].
samples were transferred to a 1 mm path-length quartz cuvette. Circular dichroism (CD) spectra in the "far UV" region (185-260 nm) was acquired with a J815 spectropolarimeter (Jasco Corp.) using a bandwidth of 1 nm and a time response of 1 s. Data collection was performed at room temperature with a scanning speed of 100 nm/min. Nine scans were obtained for each sample, and all spectra were corrected by subtracting the buffer blanks.
2.11. Measurement of PLA 2 activity and the inhibitory effect of γBjPLI upon PLA 2 PLA 2 activity was measured following the protocol described by Holzer and Mackessy [49] for 96-wells plate assay, using 4-nitro-3-octanoyloxy-benzoic acid (4N3OBA, Enzy Life Science) as substrate. The standard assay mixture contained 200 μL of buffer (10 mM Tris/HCl, 10 mM CaCl 2 , and 100 mM NaCl, pH 7.8), 20 μL of substrate, 20 μL of pure water and 20 μL of PLA 2 sample (20 μg) and were incubated for up to 40 min at 37˚C. The hydrolysis values were determined by measuring the absorbance at 405 nm at 5-min intervals. The γBjPLI inhibitory effect upon PLA 2 was determined by measuring the increase of absorbance after a previously incubation of the PLA 2 (20 μg) and γBjPLI (20 μg, 10 μg, 5 μg or 2,5 μg of γBjPLI) mixture for 20 min at 37˚C. All assays were performed in triplicate, and the absorbance at 405 nm was measured using a SpectraMax 340 multiwell plate reader (Molecular Devices). The absorbances were transformed to velocity of substrate consumption (nmol/min). The inhibition percentage was determined by linear regression through ranges of 20 to 30 minutes. The rate of substrate consumption (slope of the line) was calculated using the formula ((VoPLA2-VoγBjPLI) / VoPLA2) Ã 100.

Paw edema inhibition by γBjPLI
The γBjPLI role on edema development was analyzed by paw edema, which was induced by a single sub plantar injection in female Swiss mice (~25 g) of 10 μg of PLA 2 that was previously incubated for 30 min with 10 μg of γBjPLI (n = 5). The control groups were inoculated with 10 μg of γBjPLI (n = 5), saline (0.9%) (n = 5) or 10 μg of PLA 2 (n = 5). The paw volumes were measured immediately before the injection and at selected time intervals thereafter (0, 30, 60, 90, 120, 180, 240 and 300 minutes) using a hydroplethysmometer (model 7150, Ugo Basile). The results were expressed as the increase in paw volume (μL) calculated by subtracting the initial volume (0 min) from the final volume.

Evaluation of myonecrosis inhibition by γBjPLI
The myonecrosis inhibition by γBjPLI was determined by the injection of 20 μL of PLA 2 (20 μg) incubated for 30 min with γBjPLI (20 μg) (n = 5) on the right gastrocnemius muscle (21 g female Swiss mice). Control groups received 20 μg of PLA 2 (n = 5) or 20 μg of γBjPLI (n = 5) or 20 μL of 0.9% saline (n = 5). Mice blood samples were collected into tubes containing heparin as anticoagulant, centrifuged and the plasma was separated and stored at 4˚C for a maximum of 12 h before the assay. The amount of Creatine Kinase (CK) was then determined using 40 μL of plasma, which was incubated for 3 minutes at 37˚C with 1 mL of the reagent according to the kit protocol of CK commercial kit (Bioliquid). The resulting activity was expressed in U/L. Two-way ANOVA followed by Bonferroni post hoc test was used to analyze paw edema. The results were considered statistically significant at probability (P) values equal to or less than 0.05.

Purification of the inhibitor γBjPLI
The γBjPLI was purified by two chromatographic steps, using an ion exchange column (DEAE) (Fig 1A), followed by an affinity column (crotoxin coupled to a CNBr-activated Sepharose resin) (Fig 1B). According to the purification table the eluted fraction of affinity chromatography represents 1% of the total protein of the snake serum ( Table 1).
The electrophoretic profile of the sample eluted from the affinity chromatography was analyzed through SDS-PAGE, and showed a single band around 20 kDa in non-reduced condition, but in reduced condition this major band presented around 25 kDa, as shown in Fig 1. The γBjPLI purity was evaluated by RP-HPLC on C18 (Fig 2B).

Size exclusion chromatography protein analysis
SEC analysis showed that γBjPLI appears as a single fraction, as well as PLA 2 . Furthermore, it is possible to observe an interaction between PLA 2 and γBjPLI, when both molecules are preincubated together, through a change in the retention time profile, presenting an estimated   (Fig 2A). However, this complex cannot be observed by RP-HPLC (Fig 2B).

Mass spectrometry
Protein band corresponding to γBjPLI was excised from SDS-PAGE, digested with trypsin and analyzed by mass spectrometry, presenting 72%, 68% and 41% coverage of the amino acid sequence of proteins described as γPLI, deposited in the transcriptomic database ( Table 2). The two highest coverage proteins were identified as γPLI from B. jararaca, followed by Bothrops moojeni.

Circular dichroism analysis
γBjPLI was subjected to CD and showed the presence of many alpha-helices (26.3%) and random structures (34.5%), followed by pleated beta sheets (18.5%). In comparison, isolated PLA 2 showed approximately 24.1% of alpha-helices and 39.3% of random structures (Fig 3A).

Enzymatic assay
To confirm that the γBjPLI was able to inhibit PLA 2 , an in vitro test was initially performed. The results showed that γBjPLI was capable of inhibiting PLA 2 in a dose-dependent manner. The highest percentage was 40% of inhibition when 20 μg for γBjPLI was used (1:1 molar ratio; γBjPLI:PLA 2 ) (Fig 3B).

Paw edema and myonecrosis evaluation
After confirmation of in vitro inhibition, in vivo tests (for evaluation of edema and myonecrosis inhibition) were performed. The γBjPLI was able to significantly decrease both the edema and the myonecrotic effect of PLA 2 . In Fig 4A and 4B, it is possible to note that γBjPLI induced marginal myonecrosis as well as marginal edema, but after being preincubated with PLA 2 it showed a great decrease in the damage caused by PLA 2 .

Discussion
Although PLIs have been described and purified in several species of snakes [25,[51][52][53], B. jararaca PLI has only been reported through cDNA transcription of liver cells [32]. Therefore, this work showed for the first time, to the best of our knowledge, the isolation of a γPLI from B. jararaca serum. After two chromatographic steps, γBjPLI was isolated with a recovery of 1% of the serum proteins applied initially, which was lower than that recovered by purification of γCdcPLI (2.60%) from plasma of Crotalus durissus collineatus, possibly due to methodological differences and biological characteristics [34]. Variation in the levels of inhibitors of venom components found in snake plasmas is still not well understood, and may be a physiological response of the snakes to repeatedly contact with the venom or be under a genetically Table 2 γ-phospholipase inhibitor (γPLI) from Bothrops jararaca snake serum programmed control [54]. It is known that newborns of Clelia clelia (an ophiophagus snake), even without any contact with venom, have antihemorrhagic properties in serum [55]. On the other hand, recently, a study described the clear ontogenetic difference of inhibitors expression by qPCR profile analysis, in which γ-PLI had an up-regulation around 30-fold in adults in relation to juvenile of B. jararaca specimens [54]. Also, Kinkawa et al. demonstrated that the gene expression of α-PLI and β-PLI in Gloydius brevicaudus liver is increased by the intramuscular injection of the PLA 2 derived from its own venom [56]. γ-phospholipase inhibitor (γPLI) from Bothrops jararaca snake serum Samples of saline (20 μL), PLA 2 (20 μg), γBjPLI (20 μg) or PLA 2 + γBjPLI (20 μg + 20 μg) were preincubated for 30 min at 37˚C, injected into the gastrocnemius muscle and then the CK present in the mice blood (Swiss) was quantified and expressed in U/L. B) Neutralization of edematogenic activity of PLA 2 . Samples of PLA 2 (10 μg), γBjPLI (10 μg) or PLA 2 + γBjPLI (10 μg + 10 μg), previously incubated for 30 min at 37˚C, were injected into the right sub plantar region of the γ-phospholipase inhibitor (γPLI) from Bothrops jararaca snake serum γBjPLI was found as monomers of 20 and 25 kDa (under non-reduced and reduced conditions, respectively). This molecular mass corroborates the data described by Estevão-Costa et al. [32], obtained by primary structure deduction of transcribed γPLI in B. jararaca. However, native γPLI is usually found as oligomers formed by identical subunits of monomers, as PIP (γPLI from Asian python-Malayopython reticulatus) and BmjMIP (γPLI from B. moojeni) [31,57]. Soares et al. suggest that the oligomeric form of BmjMIP can be monomerized when occur an interaction with PLA 2 , and each subunit can bind and inactivate a myotoxin molecule [31]. So, the affinity purification process may have caused the monomerization of γBjPLI, explaining the result presented in SDS-PAGE.

. Proteins identified with higher percentage of coverage from the peptides obtained from the band excised from SDS-PAGE and analyzed by mass spectrometry in Synapt G2 HDMS (Waters
The difference found in the mass values under different conditions (with and without βmercaptoethanol) was also reported by Soares et al. [31], who suggests that the oligomers of these proteins are stabilized through non-covalent interactions, and that each subunit has internal disulfide bonds, explaining the fact that migration has been hampered by protein linearization (under reduced conditions). By contrast, under non-reduced conditions the folding is maintained, facilitating gel migration.
After preincubation of the inhibitor with PLA 2 , the interaction between the two molecules is shown by SEC, since they appear as a single fraction of about 35 kDa, which is not maintained in the RP-HPLC, suggesting that the molecular complex stability was maintained by weak molecular interaction and involves the hydrophobic interaction, easily disrupted by RP-HPLC conditions. The protein eluted by affinity chromatography was only identified after trypsin digestion, mass spectrometric analysis and comparison with the database. The γBjPLI peptide sequences obtained showed similarity with two deduced B. jararaca γPLI proteins, A8I4L6_BOTJA and A8I4M0_BOTJA, with 72% and 62% coverage, respectively. In addition, another Bothrops snake, B. moojeni, also presented a γPLI protein (A8I4N5_BOTMO) with 42% of coverage in relation to γBjPLI [32]. When sequences of γBjPLI and A8I4L6_BOTJA were aligned (Fig 5), disregarding the signal peptide sequence, we could note that the coverage increased to 85% mice paw (Swiss). The volume of edema was monitored using a hydroplethysmometer, until the decrease of the inflammation reached 20% of the initial one.
https://doi.org/10.1371/journal.pone.0193105.g004 γ-phospholipase inhibitor (γPLI) from Bothrops jararaca snake serum and that both had high similarity. The partial sequence of γBjPLI showed two substitutions: an arginine (γBjPLI) by a glycine (A8I4L6_BOTJA) at position 71, and an asparagine (γBjPLI) by a glycine (A8I4L6_BOTJA) at position 145. The region with the possible N-glycosylation site was not sequenced, probably this site should be found at position 176, since the asparagine-alanine-threonine appears in A8I4L6_BOTJA and it is a well-conserved sequence among other γPLI [32,34].
Secondary structures were analyzed by CD and showed the presence of alpha-helices (26.3%), random structures (34.5%) and beta sheets (18.5%), corroborating the analysis of Estevão-Costa [32]. Bothrops genera γPLIs have the same composition of secondary structures and the alpha-helix region was consistently described near to the carboxy-terminus (positions 157-164), according to the prediction by Estevão-Costa et al. [32].
The interaction site of γPLI and PLA 2 has been suggested to be the oligopeptide 104 QPFPGLPLSRPNGYY 118 [32], and this sequence of amino acids was identified in γBjPLI. Some authors suggest that there may be three phosphorylation sites, at position 21 (Ser), 22 (Ser) and 111 (Thr), and that these sites may help the molecule to unleash other physiological roles, besides the inactivation of PLA 2 ; however, until the moment, no other role was attributed to PLIs [32,58].
In order to evaluate the inhibition of PLA 2 activity using γBjPLI protein, it was necessary to separate the basic fraction (PLA 2 ) of crotoxin from crotapotin, which decreases the enzymatic activity of this complex [59].
Thus, isolated PLA 2 was incubated with different concentrations of γBjPLI and analyzed using the synthetic substrate 4N3OBA. Such experiment showed the inhibitory action of γBjPLI on PLA 2 enzymatic activity in a dose dependent manner. The same result had already been reported before and showed 46% inhibition of C. d. terrificus PLA 2 by γBjussuMIP, a γPLI purified from B. jararacussu plasma [58]. Since γBjPLI was able to inhibit phospholipase activity in vitro, it was decided to analyze whether it also occurred in vivo.
It is possible to note that there was an increase in plasma CK activity and paw edema caused by γBjPLI, when compared to normal levels in controls, and similar results were showed by Gimenes et al. [34]. A question that arose from this observation was if these increases in plasma CK levels and paw edema may be due to a possible contamination with crotoxin during the isolation of γBjPLI by affinity chromatography. However, chromatographic profiles obtained by size exclusion analysis and RP-HPLC showed a satisfactory purity level of γBjPLI (when applied to the column alone). In addition, it is important to point out that no peptides related to crotoxin was identified when a sample of γBjPLI was subjected to mass spectrometry analysis, confirming that there is no contamination with this molecule during the isolation of γBjPLI by affinity chromatography.
These marginal myonecrotic and edematogenic effects probably occur due to a physiological response of our experimental model, since γPLI is not a native protein of mice. However, the increase caused by PLA 2 was significantly higher than γBjPLI, enabling the evaluation of inhibitory activity, which strongly decreases when incubated with this inhibitor.
In summary, the results obtained in this work showed the isolation of a γPLI from B. jararaca serum, which can represent a new perspective for the treatment of local effects caused by Bothrops envenomation, since local effects are poorly neutralized by antibothropic serum.
Supporting information S1 Table. The complete peptide table obtained