Identification and characterization of the first endogenous phospholipase A2 inhibitor from a non-venomous tropical snake, Boa constrictor (Serpentes: Boidae)

Abstract Background: Endogenous phospholipase A2 inhibitors from snake blood (sbPLIs) have been isolated from several species around the world, with the primary function of self-protection against the action of toxic phospholipases A2. In American snakes, sbPLIs were solely described in pit vipers, in which the natural protection role is justified. In this study, we described a sbPLI in Boa constrictor (popularly known as jiboia), a non-venomous snake species from America. Methods: PLA2 inhibitory activity was tested in the blood plasma of B. constrictor using C. d. terrificus venom as the enzyme source. Antibodies developed against CNF, a sbγPLI from Crotalus durissus terrificus, were used to investigate the presence of homologues in the blood plasma of B. constrictor. A CNF-like molecule with a PLA2 inhibitory activity was purified by column chromatography. The encoding gene for the inhibitor was cloned from B. constrictor liver tissue. The DNA fragment was cloned, purified and sequenced. The deduced primary sequence of interest was aligned with known sbγPLIs from the literature. Results: The blood plasma of B. constrictor displayed PLA2 inhibitory activity. A CNF-like molecule (named BcNF) was identified and purified from the blood plasma of B. constrictor. Basic properties such as molecular mass, composing amino acids, and pI were comparable, but BcNF displayed reduced specific activity in PLA2 inhibition. BcNF showed highest identity scores (ISs) with sbγPLIs from pit vipers from Latin America (90-100%), followed by gamma inhibitors from Asian viperid (80-90%). ISs below 70% were obtained for BcNF and non-venomous species from Asia. Conclusion: A functional sbγPLI (BcNF) was described in the blood plasma of B. constrictor. BcNF displayed higher primary identity with sbγPLIs from Viperidae than to sbγPLIs from non-venomous species from Asia. The physiological role played by sbγPLIs in non-venomous snake species remains to be understood. Further investigation is needed.


Background
Secretory phospholipases A 2 are widely distributed as toxic components of snake venoms. A number of snake species express endogenous snake blood phospholipase A 2 inhibitors (sbPLIs). This kind of molecules was first described in venomous snakes with the primary function of self-protection against an eventual presence of snake venom PLA 2 (svPLA 2 ) in their own blood stream [1,2]. According to the presence of known domains from mammal proteins -C-type lectin-like, tandem leucine-rich repeats (LRRs), or three-finger motifs -sbPLIs were grouped into alpha (α), beta (β) or gamma (γ) structural classes, respectively [3]. Comparable inhibitors were later identified in a number of non-venomous species [4,5,6,7,8,9]. Whether venomous or not, some snake species express sbPLIs belonging to up to three different structural classes simultaneously [3,7,10,11].
SbγPLIs are the most widely distributed inhibitors among elapid and viperid species from the Old and New World [12,13,14]. Concerning non-venomous snakes, as far as we know, until now sbγPLIs were solely purified from Asian species [4,5,6,7,8,9]. With that in mind, we investigated the presence of this kind of inhibitor in Boa constrictor -a non-venomous tropical snake -popularly known as jiboia. We identified a functional sbγPLI, cloned the encoding gene from liver tissue and structurally characterized the deduced protein. The sbγPLI was named BcNF by analogy with CNF (Crotalus neutralizing factor), a prototype of this class of inhibitors previously isolated from the South American rattlesnake, Crotalus durissus terrificus [15,16].

Boa constrictor blood plasma and liver tissue collection
Heparinized blood plasma and liver tissue fragments were collected from a Boa constrictor specimen captured in the municipality of Contagem (19º55'54" S, 44º03'13" W), in the Brazilian state of Minas Gerais. The specimen was kept in captivity in the Serpentarium of Ezequiel Dias Foundation until death by natural causes. The whole blood was collected immediately after the animal death, centrifuged for plasma separation and clarified using a 0.22-µm microfilter. The total protein content was estimated by spectrophotometry readings at 280 nm. One optical density unit was considered to be equivalent to 1 mg/mL of protein. Liver fragments were collected in DEPCtreated tubes and quickly frozen in liquid nitrogen. Whenever applicable, blood plasma and tissue liver from C. d. terrificus specimens were used as reference.

Fractionation of B. constrictor blood plasma
Five hundred microliters of B. constrictor blood plasma were diluted to 10 mL with 25 mM Tris-HCl, 0.1 M NaCl pH 8.7 (buffer A) and dialyzed against the same buffer to ensure ionic equilibrium. After centrifugation to remove any insoluble material, the supernatant was loaded into an anion exchange column (Hitrap QFF 1mL, GE HealthCare). Protein elution was performed with a linear gradient of 25 mM Tris-HCl, pH 8.7, containing 2.0 M NaCl (buffer B), under a flow rate of 1 mL/min. Fractions with inhibitory activity (1 mL each) were pooled, 4-fold diluted with a saturated ammonium sulfate (SAS) solution and loaded into hydrophobic interaction columns connected in series [four columns HiTrap Phenyl FF 5 mL (low sub) column, GE HealthCare]. Elution was performed with a decreasing salt gradient under a flow of 5 mL/min. Total protein concentration was estimated by optical density readings of the eluted fractions at 280 nm.

Inhibition of PLA 2 activity
The crude venom of C. d. terrificus was used as a source of PLA 2 . Increasing volumes of snake blood plasma with known protein concentration were preincubated with a fixed concentration (50 μg/mL) of C. d. terrificus venom for 30 min at 37°C. The same procedure was applied to purified fractions, after dialysis against 25 mM ammonium formats, pH 6.5, whenever necessary. Residual PLA 2 activity was evaluated by measuring the clearing halos (in mm) of hydrolysis in agar gels with incorporated hen egg yolk suspension [17]. Negative (PBS) and positive (no blood plasma) controls were run in parallel. Inhibition curves were constructed by plotting the halo diameter against protein concentration in logarithm scale. Data were analyzed by linear regression using least squares method in the Graph Prism 6.0 for Mac OS X (GraphPad software Inc., California). Curve limits were calculated with 95% of confidence level. Specific activities were represented by curve slopes and expressed by mean ± S.D. Whenever applicable, regression line slopes were statistically compared in pairs.

SDS-PAGE and western blotting
B. constrictor blood plasma and purified BcNF were analyzed by SDS-PAGE in a 15% homogeneous or in an 8-25% gradient Phast® gel (Phast System®, GE HealthCare). Western blotting was revealed with rabbit anti-CNF IgG (0.5 mg/mL), followed by commercial anti-rabbit IgG-peroxidase antibody (A0545, Sigma) at a 1:5000 dilution. The color reaction was developed with DAB (3,3' diaminobenzidine tetrahydrochloride) in the presence of H 2 O 2 .

RNA extraction and cDNA synthesis
Total RNA was isolated from about 50 mg of B. constrictor liver tissue using Trizol® (Invitrogen, USA) following the manufacturer's instructions. RNA integrity was analyzed by gel electrophoresis in a 0.8% agarose gel using TBE (89 mM Tris base, 89 mM boric acid, 2 mM EDTA, pH 8.0) as running buffer. RNA bands were visualized under UV light, after staining with ethidium bromide. After cDNA synthesis using 2 to 5 µg of total RNA and oligo(dT)12-18 primer (First-Strand Synthesis kit, Invitrogen, USA), polymerase chain reactions were carried out with specific oligonucleotides based on the primary structure of CNF [15]: 3'CGCTCATGTGACTTTTGTCAC5' (sense, amino-terminus), 3'TCAGAGGCTTGCCAATCTGATG5' (antisense, carboxy-terminus). A housekeeping gene (β-actin) was PCR-amplified in parallel, in the presence of adequate oligonucleotides.
Fresh PCR products were cloned into the pGEM-T vector (Promega, USA) following the manufacturer's instructions. Insert-containing clones were isolated after PCR screening of transformed NM522 E. coli. Negative control contained no DNA. Amplified products were analyzed by electrophoresis in 1.0 % agarose gels in TBE buffer, in the presence of ethidium bromide. DNA from three positive clones were completely sequenced by the dideoxy chain termination method [18] on an automated ABI Prism 310 Genetic Analyzer (Perkin Elmer Applied Biosystems, USA) with the Big Dye Terminator Cycle Sequencing Ready Reaction (Perkin Elmer Applied Biosystems, USA). M13 forward and M13 reverse oligonucleotides were used as primers.
The cycling conditions were 3 min at 94°C, 35 cycles of 30 sec at 94°C, 30 sec at 55°C and 1 min at 72°C, followed by an extension period of 5 min at 72°C in a TC412 thermocycler (Techne).

Primary/secondary structure predictions and multiple alignment
Three complete reads in both directions were assembled and aligned against each other. The consensus sequence was used to deduce the primary structure and main basic properties of BcNF. The secondary structure was predicted using the Chou Fasman algorithm. Multiple sequencing alignments with primary structures of other sbγPLIs were performed using the ClustalW algorithm and a Gonnet's similarity matrix was subsequently generated. Inclusion criterium for sbγPLIs was the access to chemically determined or deduced primary structures in public data bases. For species with two or more sequence deposits due to isoforms, calculated consensus was taken as representative of the inhibitor. Signal peptides were removed, whenever necessary. The sbγPLIs from the following snake species were aligned: Bothrops alternatus (ABV91326/7), Bothrops erythromelas (ABV91328/9), Bothrops jararaca (ABV91330/1), Bothrops jararacussu (ABV91332/3), Bothrops moojeni (ABV91332/5), Bothrops neuwiedi (ABV91336/7), Crotalus durissus terrificus (AAA19162), Elaphe climacophora (BAH47550), Elaphe quadrivirgata (BAA83078), Gloydius brevicaudus (formerly

Identification and purification of BcNF from B. constrictor blood plasma
First, the blood plasma of B. constrictor was tested for inhibition of C. d. terrificus venom PLA 2 (Fig. 1). Inhibition was observed, although to a lesser extent when compared to C. d. terrificus blood plasma (positive control). Specific activities for PLA 2 inhibition were: -1.112 ± 0.1075 and -2.307 ± 0.1498 for B. constrictor and C. d. terrificus blood plasma, respectively. These activities were statistically different (p < 0.0001). PLA 2 inhibition activity was significantly lower for B. constrictor blood plasma.
Following, we investigated whether the observed inhibition could be due to the presence of a sbγPLIs. Western blotting revealed the presence of a CNF-like molecule in the blood plasma of B. constrictor. Two main protein bands were recognized by anti-CNF antibodies (Fig. 2), with apparent molecular masses roughly corresponding to glycosylated (ng) and non-glycosylated (ng-) monomers. A fainter band was present with mol. mass of possible dimers (2ng/2ng-). The result indicated the presence of a sbγPLIs, named BcNF, in the blood plasma of B. constrictor.
BcNF was isolated from B. constrictor blood plasma using two chromatographic steps: an ionic exchange followed by a hydrophobic interaction. The eluted fractions were assayed for PLA 2 inhibition (Fig. 3). Fractions from the second purification step presenting inhibitory activity were combined and submitted to electrophoresis and Western blotting using anti-CNF IgG.   A CNF-like molecule (BcNF) was mostly eluted with 100% of ultrapure water (Fig. 4). BcNF and CNF (positive control) at varying concentrations were assayed for PLA 2 inhibition (Fig.  5). Calculated specific activities were -1.344 ± 0.1705 and -4.797 ± 0.3434 for BcNF and CNF, respectively. These activities were statistically different (p < 0.0001). BcNF inhibited PLA 2 at a significant lesser extent compared to CNF.

BcNF cloning from B. constrictor liver tissue
The integrity of extracted RNA from B. constrictor liver tissue was confirmed by the unique presence of characteristic bands corresponding to 18S and 28S ribosomal RNAs (data not shown). After RT-PCR in the presence of specific primers for CNF, an amplicon of about 545 bp confirmed the encoding of a CNFlike protein in the liver tissue of B. constrictor (Fig. 6). The DNA fragment was cloned, purified and sequenced for further analysis.

Deduced primary structure and chemical properties predictions of BcNF
The deduced primary sequence of mature BcNF was compared to that of CNF. Both proteins are composed of 181 amino acids, including 16 conserved cysteines and a single putative N-linked   carbohydrate site at Asn 157 . Fourteen amino acid substitutions were noted in BcNF when compared to CNF, one of them (R 93 /K 93 ) within a segment proposed before for sbγPLIs interaction with PLA 2 (Fig. 7). Basic properties of BcNF and CNF are summarized in Table 1.
Amino acid substitutions, in general, lead to a decrease in the number of α-helixes from three in CNF to one in BcNF, besides a displacement of beta sheets in the predicted secondary structures of the proteins (Fig. 8).

Multiple sequence alignment of BcNF and other sbγPLIs
The deduced primary sequence of BcNF was multiply aligned with sbγPLIs from venomous and nonvenomous snakes from Asia, Australia and Latin America (available as Additional file 1). A similarity matrix was generated (available as Additional file 2) and the identity scores (ISs) were graphically represented (Fig. 9). For BcNF and sbγPLIs from Latin American pit vipers, most ISs were within the last decile (90-100%). ISs above 80% were obtained for Asian viperid snakes. On the other hand, when BcNF was compared to sbγPLIs from non-venomous species from Asia, the ISs were below 70%. ISs below 70% were also obtained for Elapidae snakes.

Discussion
Boa is a Neotropical genus of snakes that occurs almost continuously from southern South America through to northern Mexico [19]. Historically recognized as monotypic, recent data based on the distinct morphological traits, color patterns exhibited by these snakes and the wide diversity of ecosystems they inhabit, collectively suggest that the genus contains multiple species [20]. In Brazil, B. constrictor (sensu lato) can be found all over the country, except in the extreme south [21]. It is an aglyphous species, devoid of venom or Duvernoy's glands.
Similarly to other henophidian snakes (boas, pythons and their kin), B. constrictor uses constriction to subdue and kill a wide range of prey -including lizards, birds and mammals -through an interesting modulated process mediated by the victim's heartbeat [22]. Apparently, there is no need of an inhibitor for self-protection against toxic svPLA 2 .
B. constrictor inhibition of PLA 2 was lower than that of C. d terrificus blood plasma. Similarly, BcNF was less active than CNF. Our results are in accordance with those described for E. climacophora and E. quadrivirgata. Respective sbγPLIs were detected at higher amounts in the former, and justified by the ophiophagous habits of the species [7]. It is important to note that, in addition to sbγPLIs, those Elaphe species express sbα− and sbβPLIs simultaneously in the circulating blood. We used antibodies developed against CNF to search for sbγPLI in B. constrictor. The detection of inhibitors from other structural classes is a possibility that cannot be discarded.
BcNF is highly similar to CNF, with 14 substitutions in a total of 181 amino acids and an IS of about 90%. The molecular masses of the non-glycosylated monomers, calculated from amino acid compositions, are very close (Table 1). Band migrations in gel electrophoresis also indicated similar apparent molecular masses for monomers and oligomers (Fig. 7). Like CNF, BcNF is composed by a mixture of non-glycosylated (20 kDa) and glycosylated (22-24 kDa) monomers. For CNF, which is the main subject of study in our lab, the proportion between non-glycosylated and glycosylated varies according to the preparation. The sample loaded in SDS-PAGE (Fig. 2) was mostly nonglycosylated. However, it has been shown that the carbohydrate moiety is not essential for PLA 2 inhibition by CNF [24]. The same might be true for BcNF. The tendency for oligomerization might be a shared property, too. In fact, the 16 th , 113 th , 132 nd and 166 th tyrosinyl residues, which were previously suggested to form the interface between monomers in the oligomerization of CNF, are maintained at the same positions in BcNF. These residues might be involved in the oligomerization of the latter also. BcNF was only tested against svPLA 2 from C. d. terrificus venom, but the possibility of inhibition of other svPLA 2 cannot be discarded. The decapentapeptide Q 84 PFPGLPLSRPNGYY 98 , which was previously proposed to be the best consensus motif possibly involved in the sbγPLIs interaction with PLA 2s is maintained in BcNF. The only amino acid replacement was conservative (R 93 /K 93 ).
Interestingly, BcNF appeared more closely related to sbγPLIs from Latin American pit vipers, and from Asian pit vipers to a lesser extent, than to those from non-venomous snakes from Asia described so far.

Conclusion
A functional sbγPLI (BcNF) was described, for the first time, in the blood plasma of B. constrictor, a non-venomous species from America. BcNF displayed higher primary identity with sbγPLIs from pit vipers than with sbγPLIs from non-venomous species from Asia. Even with a growing number of sbγPLI identifications in the last years, the physiological role played by these proteins in non-venomous snake species remains to be clarified.