Six fractions obtained by fractionating B. asper venom on ion exchange chromatography on CM-Sephadex C-25 were evaluated for PLA₂ activity. It was found that fraction V was the only positive fraction for PLA₂ activity (see Figure 1A). However, fraction VI, corresponding to a PLA₂ homologue devoid of enzymatic activity (see Section 3.2), was also analyzed for antiplasmodial activity to determine the possibility of catalytically-independent actions. Fractions V and VI were subjected to further separation by RP-HPLC on a C₁₈ column. This separation revealed that fraction V had four subfractions (see Figure 1B,C), only one of which (V-4) showed PLA₂ activity, whereas fraction VI showed only one peak. These two fractions were used to assess antiplasmodial activity.

Fraction V had a minimal indirect hemolytic dose of 1.35 µg, while fraction VI showed no such activity. The PLA₂ isolated by HPLC from fraction V showed a minimum indirect hemolytic dose of 0.82 µg, while the peak obtained by HPLC separation of fraction VI lacked activity (data not shown). The hemolysis test with different substrates (egg yolk, plasma or human serum) yielded similar results in all assays. When indirect hemolytic activity was determined in solution, 100% hemolysis was observed using concentrations of 25 µg/mL for whole venom and 12.5 µg/mL for fraction V, whereas fraction VI lacked PLA₂ activity in all tests (see Figure 2).

Both venom and fractions V and VI exhibit antiplasmodial activity on the FCB1 strain of P. falciparum, with fraction V being more active than fraction VI (see Table 1). On the other hand, the venom was more active than the two fractions evaluated. Guillaume et al. showed that removal of phospholipids from cultures of P. falciparum reduced the antiplasmodial activity of PLA₂ [27], confirming the crucial role of PLA₂ enzymatic activity to control the growth of parasites in this test. Our data demonstrate the antimalarial efficacy of fraction with PLA₂ activity. However, a PLA₂ homologue devoid of enzymatic activity also resulted in restriction of P. falciparum multiplication, confirming a catalytically-independent antiplasmodial activity. This effect could be due to the perturbing action exerted by the PLA₂ homologue in the plasma membrane, thus resulting in an increase in permeability [29]. It has been shown that the C-terminal region of these PLA₂ homologues is responsible for this catalytically-independent membrane perturbation, as demonstrated in bacteria [16,30,31], being, therefore, a different mechanism from the one described for other PLA₂s [26,27].

The changes observed in the intraerythrocytic development of Plasmodium indicate that structural changes occur, as well as modifications in membrane functions in parasitized red blood cells. In addition, increments and changes in the permeability of the membrane have been described, together with the appearance of new parasite-derived proteins and changes in the composition of membrane lipids [32,33]. The observed increased permeability could also be responsible for the PLA₂ activity on the parasite, as demonstrated by Moll et al., who noted that in the absence of serum in the culture in vitro, PLA₂ lysed parasitized cells [34]. This increase in membrane permeability could also enhance the antimalarial activity of the PLA₂ homologue observed in our experiments.

Electrophoresis showed that proteins of fractions V and VI (lanes 2 and 4, respectively) had molecular weights ranging from 25 kDa and 14 kDa, when fractions were separated in non-reducing conditions, thus evidencing the presence of monomers and dimers, whereas only bands of around 14 kDa were observed (lanes 3 and 5 in Figure 1D, respectively) when these fractions were subjected to reducing conditions, thus corresponding to PLA₂ monomers (Figure 1D).

We determined the molecular mass of each of the fractions obtained by RP-HPLC: Fraction V (fractions V-1, V-2, V-3 and V-4) and VI. Mass spectrometric analysis showed that V-1 was of 13786.9 Da, V-2 was of 13950.1 Da, V-3 was of 13972.4 Da, V-4 was of 13974.6 Da and VI was of 13725 Da. The tandem mass MS/MS analysis indicated that the PLA₂s isolated corresponding to the fractions V-1, V-2, V-3 and VI were K49 PLA₂ homologs, and V-4 was D49 PLA2 (Table 2).

Additionally, the identified peptides were subjected to BLAST analysis to determine their identity with other phospholipases. The results confirmed the high identity of these peptides with PLA₂s from the venoms of B. asper, B. neuwiedi, B. jararacussu, B. pirajai and Cerrophidion godmani, among others (see Figure 3, Figure 4, Figure 5, Figure 6, Figure 7).

The results of the alignments show that the PLA₂s and PLA₂ homologues purified from the venom of B. asper from Colombia are similar to other PLA₂s and PLA₂ homologues present in other Bothrops snakes. In addition, the PLA₂ D49 shows homology with other PLA₂s from Bothrops, being higher with those of B. asper from Costa Rica (see Figure 5).

Analysis of the cytotoxic effect of the whole venom and the different fractions tested showed that fraction V was more cytotoxic than whole venom or fraction VI on PBMC cells (see Figure 8).

The cytotoxic activity of venoms and PLA₂s is a problem in using these in future biomedical applications. However, our results show that the PLA₂ isolated exerts an antimalarial effect at a lower dose than that required to induce cytotoxicity in PBMC and indirect hemolysis.

Other authors have shown that cytotoxic activity is dependent on serum in suspensions of tumor cells and red blood cells [43]. In some experiments, we cultured cells with fetal bovine serum 2% (FBS) and inactivated serum or plasma, and in these conditions, the cytotoxic dose was still higher than the antimalarial dose (results not shown).

The LD₅₀ of the whole venom of B. asper was 3566 µg/kg (2561 to 3693), whereas no lethality was observed in mice injected with fractions V and VI at doses as high as 15,000 µg/kg (see Table 1).

The envenoming of B. asper induces local and systemic symptoms, such as edema, pain and bleeding, among others, due to the effect of different toxins in the venom, such as PLA₂, serine proteinases and metalloproteases, among others [19,44,45,46,47,48,49]. The low toxicity of fraction V and of the PLA₂ homologue isolated from fraction VI compared with the venom indicates their low overall toxicity in mice and reinforces the concept that these fractions are good lead compounds in the search for antimalarial activity. This is in agreement with reports on the use of snake venom PLA₂s to inhibit microorganisms, such as bacteria and fungi, as well as parasites including Giardia duodenalis, Trypanosoma cruzi, Leishmania spp and P. falciparum [17,30,31,50,51,52].

The venom was obtained by manual milking of 40 adult specimens from different regions of Colombia held in captivity at the Serpentarium of the University of Antioquia (Medellín, Colombia). Once extracted and pooled, the venoms were centrifuged (3000 rpm, 15 min), and the resulting supernatants were lyophilized and stored at −20 °C until use.

Acetonitrile (CH₃CN) and trifluoroacetic acid (CF₃COOH) HPLC grade were purchased from Fisher Scientific (Loughborough, UK). Histopaque^®-1077, RPMI-1640 medium culture, Thiazolyl Blue Tretrazolium Bromide (MTT) and dimethyl sulfoxide (DMSO) were purchased from Sigma (Sigma-Aldrich, St Louis, MO, USA). Water for HPLC was deionized to a degree of purity of 17 Ω.

PLA_2s were purified from 50 mg of whole venom of B. asper dissolved in phosphate-buffered saline (PBS), pH 7.2, and passed through a CM-Sephadex C₂₅ ion exchange column (1.8 × 120 cm) at the flow rate 1.0 mL/min on a low-pressure chromatography system (Econo-System, BioRad, Hercules, CA, USA). The resulting fractions were analyzed for their PLA₂ activity and then PLA₂ positive fractions submitted to a reverse phase HPLC (RP-HPLC) (Shimadzu, Model Prominence, Shimadzu Corporation, Kyoto, Japan) in a C₁₈ column (pore 5 µm, 250 mm × 4.6 mm mark RESTEK Bellefonte, PA, USA) using a linear gradient (0%–100%) acetonitrile (v/v) in 0.1% (v/v) trifluoroacetic acid at a flow rate 1.0 mL/min. Finally, fractions were lyophilized and stored at −20 °C until use.

Protein homogeneity of the obtained fractions were determined by electrophoresis under reducing and non-reducing conditions in SDS-polyacrylamide gel electrophoresis (SDS-PAGE) 15% [53]. Protein molecular weight was estimated according to a molecular weight markers range of 97.4 to 14.4 kDa (BioRad, Philadelphia, PA, USA). The gels were stained with Coomassie Brilliant Blue G-250. The molecular masses of PLA₂ fractions were confirmed by direct-infusion mass spectrometry in an IonTrap (series 6310, Agilent Technologies, Santa Clara, CA, USA).

The PLA₂s and PLA₂ homologues isolated from B. asper venom (fractions V and VI see results, Figure 1B,C) were digested in solution with trypsin (0.1 ng) at 30 °C (Agilent Technologies, Santa Clara, CA, USA) overnight, according to the manufacturer’s instructions, and injected onto a nano LC-ESI-MS/MS system (Agilent Technologies, Santa Clara, CA, USA) using a nano column C₁₈ (Agilent Zorbax 300SB-C18, 150 × 0.075mm, 3.5 μm) coupled to a mass spectrometer IonTrap MSD series 6300 (Model 6310, Agilent Technologies, Santa Clara, CA, USA) [54]. MS/MS mass spectra were obtained in positive mode, dynamic range 200–1200 Da; Electrospray at 2 kV and 230 °C dry temperature, trap drive 200 ms. Charged state deconvolution of the MS/MS spectra were determined using the ChemStation G2070-91126 (Agilent Technologies, Santa Clara, CA, USA).

Deconvoluted profile spectra were used to search online the MASCOT [55] and Spectrum Mill (Agilent Technology, Santa Clara, CA, USA) in the NCBInr database for protein identification. The parameters of the search included digestion with trypsin and Carbamidomethyilation modified (C) as fixed modification. The minimum score for the intensity of each fraction was 50%, monoisotopic mass, mass tolerance of 2.5 Da and a way to search for identity.

The identified peptides were subjected to a BLAST search [56] to determine the homology with other PLA₂ family proteins. This homology was performed in BLASTP, the search parameters being non-redundant protein sequence (nr) and a snake organism.

The Median Lethal Dose (LD₅₀) was determined by the Spearman-Karber method (World Health Organization, 1981) using groups of four mice (Swiss-Webster mice strain) injected intraperitoneally (IP) with varying doses of either fractions or whole B. asper venom, previously dissolved in 0.5 mL PBS, pH 7.2. Fatalities were recorded within 48 h, and the results were expressed as the average of three repetitions.

Peripheral blood mononuclear cells (PBMC) were separated by centrifugation of citrated human blood (400g, 30 min) with Histopaque^®-1077 (Sigma-Aldrich, St Louis, MO, USA), washed with PBS and transferred to 96 well plates at a concentration of 3 × 10⁵ cells/well. Cells were cultured with different concentrations of fractions (37 °C, 5% CO₂) for 24 h. After this time, 40 µL of MTT was added and incubated for 3 h (same conditions as described). The reaction was halted by adding 130 µL of dimethyl sulfoxide (DMSO) and readings were performed in a microplate reader at 420 nm. The 50% cytotoxic dose was calculated by linear regression [57].

This was evaluated following the method that uses agarose gel-erythrocyte-egg yolk [58,59]. We estimated the minimum indirect hemolytic dose (MIHD), defined as the dose of venom producing a hemolytic halo of 20 mm in diameter after 20 h. In addition, indirect hemolytic activity was assessed on red blood cells in suspension. For this, different doses of either the whole venom or fractions V and VI were incubated with fresh human red blood cells for 30 min at 37 °C in the presence of 250 µL of inactivated human serum, inactivated human plasma, egg yolk or PBS. Afterwards, samples were centrifuged, and the percentage of lysis was determined by recording the absorbance at 540 nm as an index of released hemoglobin. As a control of 100% hemolysis, 2%Triton X-100 was used. The results were expressed as percentage of lysis, and the venom or toxin concentration producing 100% hemolysis was determined.

Based on the procedure described by Trager and Jensen [60], parasites were grown at 37 °C in A+ human erythrocytes to a hematocrit of 2% and 3%–6% parasitemia under an atmosphere of 3% CO₂, 6% O₂ and 91% N₂.

Increasing concentrations of PLA₂ fractions V and VI in complete medium were plated in 96-well plates (100 µL/well) and incubated with asynchronous P. falciparum FCB1 (1.5% parasitemia, 4% hematocrit, 100 µL/well). Parasites were incubated as previously described [60]. After 24 h, 0.5 mCi of ³H-hypoxanthine was added to the culture, and parasites were cultured for further 24 h at the same conditions. Finally, the plates were freeze-thawed, and parasites were harvested onto filter paper, added to liquid scintillation cocktail and the incorporation of ³H-hypoxanthine determined in a Microbeta counter 1450 (Wallac, Perkin Elmer, Waltham, MA, USA).

The percentage of growth inhibition was calculated based on 100% uptake of the 3H-hypoxanthine of controls (parasites in culture medium, incomplete RPMI). Growth inhibition was calculated based on 100% uptake of the 3H-hypoxanthine control in parasites in the absence of PLA₂s or PLA₂ homologues. The IC₅₀ values correspond to the venom or toxin concentration required to kill 50% of the parasites within 48 h, and was determined from dose-response curves according to Desjardins et al. [58].

The results are presented as mean ± S.E.M of three replicates, and experimental differences between means were determined by analysis of variance followed by Dunnett’s test for intragroup comparisons. Significance was set up at p < 0.05.