Venom complexity of Bothrops atrox (common lancehead) siblings

Abstract Background: Variability in snake venoms is a well-studied phenomenon. However, sex-based variation of Bothrops atrox snake venom using siblings is poorly investigated. Bothrops atrox is responsible for the majority of snakebite accidents in the Brazilian Amazon region. Differences in the venom composition of Bothrops genus have been linked to several factors such as ontogeny, geographical distribution, prey preferences and sex. Thus, in the current study, venom samples of Bothrops atrox male and female siblings were analyzed in order to compare their biochemical and biological characteristics. Methods: Venoms were collected from five females and four males born from a snake captured from the wild in São Bento (Maranhão, Brazil), and kept in the Laboratory of Herpetology of Butantan Intitute. The venoms were analyzed individually and as a pool of each gender. The assays consisted in protein quantification, 1-DE, mass spectrometry, proteolytic, phospholipase A2, L-amino acid oxidase activities, minimum coagulant dose upon plasma, minimum hemorrhagic dose and lethal dose 50%. Results: Electrophoretic profiles of male’s and female’s venom pools were quite similar, with minor sex-based variation. Male venom showed higher LAAO, PLA2 and hemorrhagic activities, while female venom showed higher coagulant activity. On the other hand, the proteolytic activities did not show statistical differences between pools, although some individual variations were observed. Meanwhile, proteomic profile revealed 112 different protein compounds; of which 105 were common proteins of female’s and male’s venom pools and seven were unique to females. Despite individual variations, lethality of both pools showed similar values. Conclusion: Although differences between female and male venoms were observed, our results show that individual variations are significant even between siblings, highlighting that biological activities of venoms and its composition are influenced by other factors beyond gender.


Background
Snakebite envenomation is considered a worldwide Category A neglected tropical disease and constitutes a public health problem in warmer regions of the developing world [1,2].In Latin America, the family Viperidae is responsible for most of the registered snakebite accidents, and in Brazil, the genus Bothrops is responsible for 85% of the ophidian envenomation [1][2][3][4][5].
Bothrops atrox (common lancehead) is a pit viper species widely distributed in the northern region of South America [7][8][9] and its natural history is already well documented [10].This generalist species occurs mostly in rainforests, but can also be found in disturbed areas.In relation to other Bothrops species, the common lancehead shows preference towards heavier preys [11].Males are smaller than females and are more prone to higher mortality, considering the active foraging lifestyle of the species.In fact, B. atrox exhibits a dynamic use of its habitat, being known as one of the most active hunters of the Bothrops genus [9,11,12].B. atrox venom causes mainly local damage, such as edema, hemorrhage and necrosis, apart from systemic effects, including blood coagulation disorders [13,14].In lethal cases, hemorrhage leads to cardiovascular shock and acute renal failure secondary to acute tubular necrosis and occasionally glomerulonephritis [7,15].These symptoms are the result of individual or synergistic action of different toxins that comprise the venom of snakes [16,17], such as phospholipases A 2 (PLA 2 s), metalloproteinases (SVMPs), serine proteinases (SVSPs), L-amino acid oxidases (LAAOs), among others [1,18].The knowledge about the composition and action of snake venoms allows us to understand the evolutionary processes in ophidians [19] and elucidate the mode of action of toxins and the demand for their antagonists [20].In addition, as snake venoms are a rich source of bioactive compounds with pharmaceutical potential, they can represent an improvement in snakebite envenoming treatment, which can impact significantly on the victims symptoms and the quality and efficacy of antivenoms [21,22].
Individual variability is a well-established concept when referring to intraspecific variation of snake venom composition and/or its activities, and may be related to ontogeny [23][24][25], diet [26,27], seasonality [28], geographical location [29][30][31], gender [32][33][34][35], and captivity [22,36].Within the Bothrops species, B. jararaca venom is the most studied one regarding gender differences [32,37], contrary from B. atrox, despite its high geographic distribution and epidemiological representation.In this context, the present study aims to compare, for the first time, the biochemical and biological characteristics of male and female venom of B. atrox siblings.Both genders were born in captivity and maintained under controlled conditions, in order to contribute to the knowledge of changes in venom characteristics according to sex, as well as the formulation of pharmacological tools for inhibiting the toxic effects of this venom.

Animals
Mus musculus (Swiss) male mice (18-22 g) were obtained from Butantan Institute animal house, had access to water and food ad libitum and were kept under a 12 h light/dark cycle.B. atrox specimens (5 females and 4 males over 11 years of age) (Additional file 1) were born from the same snake captured from the wild (São Bento, Maranhão, Brazil), and kept in the Laboratory of Herpetology of Butantan Institute under controlled conditions.

Venoms
The venom was extracted from nine B. atrox snakes (5 females and 4 males born from the same mother), centrifuged for 15 min at 1700 × g, 4 ºC, to remove any scales or mucus, lyophilized, and stored at -20 ºC until use.Information regarding the snakes is available in Additional file 1.

Protein quantification
Protein concentration of pools (female and male) and individual venom samples was determined according to the Bradford method, using Bio-Rad Protein Assay reagent and bovine serum albumin (BSA) (Sigma) as standard [38].These data were only used as a basis to other experiments.

One-dimensional electrophoresis (1-DE)
Electrophoretic analysis of pools and individual venom samples was performed using 30 µg of protein in the presence and absence of β-mercaptoethanol in 15% polyacrylamide gels [39].The gels were stained with Coomassie Blue G according to the GE Healthcare protocol.

Protein identification by mass spectrometry
Identification of proteins was performed by LC-MS/MS in a Synapt G2 (Waters) coupled to the nanoAcquity UPLC chromatographic system (Waters) as previously described [40,41].Briefly, samples of 100 μg of protein from each venom pool were incubated in 50 mM ammonium bicarbonate with 5 mM DTT (dithiothreitol) for 25 min at room temperature (RT), followed by addition of 14 mM IAA (iodoacetamide) and incubation in the dark for 30 min at RT. Finally, an incubation with 5 mM DTT for 15 min was performed.Calcium chloride (1 mM) and 1 µg of trypsin (Sigma) in 50 mM ammonium bicarbonate were added to each sample and incubated for 16 h at 37 °C.After incubation, the reaction was stopped with 5% TFA (0.5% final concentration).Aliquots of the resulting peptide mixtures (5 μg) were injected into a trap column packed with C18 (nanoAcquity trap Symmetry 180 μm × 20 mm) at 8 µL/min with phase A (0.1% formic acid.Peptides were then eluted onto an analytical C18 column (nanoAcquity BEH 75 μm × 200 mm, 1.7 m) at a flow rate of 275 nL/min, using a gradient of 7-35% of phase B (0.1% formic acid in acetonitrile) in 90 min.Data were acquired in the in data-independent mode UDMSE [42] in the m/z range of 50-2000 and in resolution mode.Collision energies were alternated between 4 eV and a ramp of 17-60 eV for precursor ion and fragment ions, respectively, using scan times of 1.25 s.The ESI source was operated in positive mode with a capillary voltage of 3.0 kV, block temperature of 70 °C, and cone voltage of 40 V.For lock mass correction, [Glu1]-Fibrinopeptide B solution (500 fmol/mL in 50% acetonitrile, 0.1% formic acid; Peptide 2.0) was infused through the reference sprayer at 500 nL/min and sampled for 0.5 s at each 60 s.
Raw data were processed in ProteinLynx Global Server 3.0.1 (Waters) by the Apex3D module using low energy threshold of 750 counts and elevated energy threshold of 50 counts.MS/ MS spectra were submitted to searches a Serpentes database (downloaded from Uniprot in March 1 st , 2019, 2608 reviewed sequences).The following search parameters were used: automatic fragment and peptide mass tolerances, carbamidomethylation of cysteines as fixed modification, oxidation of methionine, N-terminal acetylation, glutamine and asparagine deamidation as variable modifications, up to 2 missed cleavage sites were allowed for trypsin digestion.The following criteria were set for protein identification: a minimum of 1 fragment ion per peptide, 5 fragment ions per protein and 2 peptides per protein, and a maximum false discovery identification rate of 1%, estimated by a simultaneous search against a reversed database.Label-free quantitative assessments were based on the average intensities of the three most intense peptides of each identified protein [43].Each pooled sample was analyzed in technical triplicate.Data of the spectra are available in Additional file 2.

Caseinolytic activity
Caseinolytic activity was determined as described [44] using azocasein (Merck) as substrate.Briefly, 85 µL of a 4.25 mg/mL azocasein solution were incubated with 10 µL of each venom (1 mg/mL), both diluted in 50 mM Tris-HCl buffer, pH 8.0.The reaction was stopped by adding 200 µL of 5% trichloroacetic acid (TCA).The samples were centrifuged at 1000 × g and 100 µL of the supernatant were homogenized with 100 µL of 0.5 M NaOH.The absorbance was measured at 450 nm in a SpectraMax i3 microplate reader (Molecular Devices).One unit of activity was determined as the amount of venom that induces an increase of 0.005 units of absorbance.

Collagenolytic activity
Collagenolytic activity over azocoll was determined according to Váchová and Moravcová [45] and modified by Antunes et al. [46].Venoms (6.25 µg) were incubated with 50 µL of a 5 mg/ mL azocoll (Sigma) solution, both diluted in Tyrode buffer (137 mM NaCl, 2.7 mM KCl, 3 mM NaH 2 PO 4 , 10 mM HEPES, 5.6 mM dextrose, 1 mM MgCl 2 , 2 mM CaCl 2 , pH 7.4) for 1 h in constant shake, at 37 °C.The samples were centrifuged for 3 min at 5000 × g and the absorbance of the supernatants (200 µL) was measured at 540 nm in a SpectraMax i3 microplate reader (Molecular Devices).One unit of activity was determined as the amount of venom that induces an increase of 0.003 units of absorbance.

L-amino acid oxidase activity
Pools and individual venom samples were analyzed by measuring the hydrogen peroxide generated during the oxidation of L-amino acids [47].For this, 5 μg of the venom were added to the 90 µL reaction mixture containing 50 mM Tris-HCl, 250 mM L-methionine, pH 8.0, 2 mM o-phenylenediamine and 0.8 U/mL of horseradish peroxidase, and the mixture incubated at 37 ºC for 60 min.The reaction was stopped using 50 μL of 2 M H 2 SO 4 and the absorbance measured on a spectrophotometer (SpectraMax i3, Molecular Devices) at 492 nm.Results were expressed as 1 μM of H 2 O 2 /minute/µg of venom.

Phospholipase A 2 activity
The phospholipase A 2 (PLA 2 ) activity of pools and individual venom samples was determined based on the assay developed by Holzer and Mackessy [48] using the monodisperse synthetic substrate 4-nitro-3-octanoyloxy-benzoic acid (NOBA).Twenty µg of venom (dissolved in 0.85% NaCl), 20 µL of deionized water and 200 µL of 10mM Tris-HCl, 10 mM CaCl 2 , 100 mM NaCl, pH 8.0 were mixed in a 96 well microplate.Then, 20 µL of NOBA (4.16 mM in acetonitrile) was added in a final concentration of 0.32 mM.After incubating for 20 min at 37 ºC, the absorbance at 425 nm was recorded in a microplate reader (SpectraMax i3, Molecular Devices).A change of 0.1 absorbance unit at 425 nm was equivalent to 25.8 nmoles of chromophore release.

Coagulant activity
The coagulant activity of the venom pools was assessed in citrated human plasma, according to Theakston and Reid [49].Briefly, 100 μL of plasma were incubated at 37 °C for 60 s.After the incubation, 50 µL of various concentrations of venom samples were mixed and clotting times were measured in a coagulometer (MaxCoag, MEDMAX).The Minimum Coagulant Dose (MCD) was defined as the minimum amount of venom that induced coagulation of plasma in 60 s at 37 °C.

Hemorrhagic activity
The hemorrhagic activity was obtained by the determination of Minimum Hemorrhagic Dose (MHD).Groups of five male Swiss mice of 18-22 g were injected with 100 μL of several doses of venom pool samples, diluted in 0.89% NaCl, intradermally into the venter of the mice, and a control group received 100 μL of NaCl solution under identical conditions.After 3 h, the animals were euthanized in a CO 2 chamber, the venter skin was removed, and the hemorrhagic areas were measured [50].The MHD was defined as the amount of venom that produced hemorrhages with a mean diameter of 10 mm after 3 h [51].

Median lethal dose (LD 50 )
The LD 50 of venom pool samples were determined by intraperitoneal injection in 18-22 g male Swiss mice with 500 μL of varying doses of venoms (66-381 µg/animal) in 0.89% NaCl.Five mice were used per group and the number of deaths occurring within 48 h after injection was recorded.The LD 50 and 95% confidence intervals were calculated by Probit analysis [52].

Statistical analysis
Results are expressed as mean ± SD of triplicates.The significance of differences between the means of the venoms was determined by one-way ANOVA with Tukey as a posteriori test and venom pools were analyzed using Student's t-test using GraphPad Prism 7.03 software, where p < 0.05 was considered significant.

Results and Discussion
Differences in the composition and activity of snake venoms from the same species are a worldwide researchers concern.These differences can influence directly in the antivenom production and in the success of patient treatment [54][55][56][57].

Compositional analysis
Although B. atrox venom has been analyzed in several aspects [30,[58][59][60], this work showed, for the first time, a comparative study of the venom extracted from female and male siblings, born in captivity and kept under controlled environmental conditions.
Electrophoretic profiles were evaluated, showing similar band patterns with few differences between individuals and pools.Individual analysis of non-reduced venoms showed a common band of ~35 kDa (Figure 1A), which is only present in the venoms of females and of Ba8 among males, and another band of ~30 kDa that is present only in the venom of males, except for Ba8.These two bands might be associated to P-II SVMP and SVSP respectively, in accordance with their molecular masses [24], and their presence and absence are reflected in the pool, although faint (especially the ~35 kDa band).Moreover, it is possible to observe bands of less intensity between 25-50 kDa (probably CRISP, GPC, P-I and P-III SVMP and SVSP) and over 100 kDa (most likely PDE).These results have been observed not only in B. atrox but also in other snakes of the Bothrops genus, and are supported by several works [31,[61][62][63].
In order to compare the composition of female and male B. atrox venoms, they were pooled according to gender and submitted to in-solution trypsin digestion followed by LC- MS/MS analysis on a Synapt G2 mass spectrometer (Waters).The results obtained allowed to identify 112 different protein compounds (Table 1 and Additional file 3), of which 105 were common proteins between female and male venom pools and 7 were unique to females.Proteins identified belong to the following families: SVMPs, SVSPs, LAAOs, CTLs, PLA 2 s, nucleotidase (NT), phospholipase B (PLB), glutaminyl-peptide cyclotransferases (GPCs), cysteine-rich secretory protein (CRISP), and disintegrin-like protein (DISL) (Figure 2, Table 1 and Additional file 3); the first five families are the main    compounds in Bothrops venoms [32,[64][65][66].The unique proteins identified in the female venom were one LAAO, one P-I SVMP, one P-III SVMPs, one DISL, one CRISP, and two fragments of SVSPs.The Bpic-LAAO is a high weight protein of 65 kDa that causes edema and inhibition of platelet aggregation [67]; the P-I SVMP (barnettlysin-1) is non-hemorrhagic and is known to cleave many substrates, including fibrin(ogen), but not collagen [68]; VAP-1 is a P-III SVMP related to hemorrhagic activity, but is unable to cleave collagen [69]; leberagin-C is a DISL that inhibits platelet aggregation [70]; the exclusive CRISP found in the female venom was catrin-2, which weakly blocks muscle contraction induced by K + and Ca 2+ channels [71].Sousa et al. [30] examined the venom composition of B. atrox according to their habitats and the proteomics analyses showed some differences in comparison to our study, such as the presence of hyaluronidases, which were not identified in this work.It is interesting to note that the relative percentages of LAAOs and SVSPs obtained by our group by MS analysis were higher than the aforementioned study, 16% in comparison to ~9% for LAAOs, and 21% in comparison to 10% to 14% for SVSPs, respectively.Another study indicates higher percentages of SVMPs than found here and have not detected any PLB [60].

Functional analysis
Proteolytic activities over casein and collagen did not show statistical difference between female and male pools, although some individual variations were observed.For caseinolytic activity (Figure 3A), only Ba4 and Ba6 showed statistical difference.As for collagenolytic activity (Figure 3B), individual variability was more evident.Caseinolytic activity may be associated with SVMP and SVSP, since casein is a substrate degraded by these families of proteins [72,73] and, in this study, neither of these two protein families differed between the pools analyzed by MS (Figure 2).
LAAOs have the ability to induce or inhibit platelet aggregation, in addition to promoting hemorrhage, hemolysis, the appearance of edema, and other biological activities [74][75][76].The percentage of LAAOs found in female venom pool analyzed by MS was slightly higher than for males.However, male venom pool showed higher activity compared to the female pool (Figure 3C).Although contrasting, the same behavior was observed in B. moojeni [34].Similar to collagenolytic activity, LAAO activity differed individually.
PLA 2 activity (Figure 3D) of B. atrox venom showed a strong individual variation, but, overall, the venom of males presented higher activity than female venoms.This was also reflected in the pools: male pool had a higher activity than female pool.Similar results were also observed in other species, like B. jararaca and B. moojeni [34,77].This result was corroborated by mass spectrometry identification, in which a higher percentage of PLA 2 was found in the male pool.In Viperidae, the PLA 2 s found in snake venoms have been divided into two groups: with catalytic activity (Asp49 -D49) and without catalytic activity (Lys49 -K49).The substitution of the amino acid residue Asp-49 for Lys-49 consequently causes loss of calcium binding, primordial for its enzymatic activity [78].
In MCD analysis (Figure 3E), female venoms showed very similar activity among them, as well as the pool.As for males, Ba8 showed the highest activity, comparable to females, while the others presented much lower activity in comparison to females.The MCD is most likely attributed to procoagulant SVMPs and SVSPs, relating to activation of prothrombin and factor X of the clotting cascade [79,80].Despite similarities in abundance between the groups, the female pool showed, altogether, slightly more SVSP than male pool in proteomic analysis.Besides, female venom pool had slightly higher amount of thrombin-like than the male pool (11.0% and 10.6%, respectively) (Figure 2, Table 1 and Additional file 3).Also, if we consider that 112 proteins were identified in the mass spectrometry of B. atrox snake venoms used in this study and that each protein-protein interaction responds differently depending on the compounds involved [16,17], this difference may also be attributed to the synergy between protein families in local and systemic damage.It is important to highlight the limitations of the use of plasma without recalcification in this work because this may influence the time of clotting of each venom.Although it is known that SVMPs from the group A are not dependant of cofactors (including calcium) to activate prothrombin [81], a recent study [82] showed that the procoagulant effects of Bothrops genus snake venoms are highly dependant of calcium and that the dependency varies between populations.Although the results obtained herein show that, in the absence of calcium, the venom of females B. atrox is prone to be more coagulant, it is important to consider the role of calcium upon snake venom coagulopaties, even for independent calcium prothrombin activators [83], which may result in a misinterpretation of the relative toxicities.
Individual differences were observed in enzymatic activities, highlighting the importance of individual analysis when possible.Despite some individual differences, a pattern between the activities of females and males can be correlated, so, for in vivo tests, the pool was chosen for analysis.Galizio et al. [84] reinforce the importance of the individual analysis, but for ethical issues pools were used to reduce the number of animals utilized in the in vivo experiments.
MHD of male venoms was lower when compared to females (2.7 and 4.8 μg/animal, respectively), indicating that female venom pool needs more than 43.8% of venom to generate the corresponding hemorrhagic halo to MHD, than male venom pool.Saldarriaga et al. [51] found 1.8 µg/animal as MHD for adult (3 years old) B. atrox, a minor dose than the one found in this work.Although considered adults, these snakes were younger than the ones in our work.Guércio et al. [24] analyzed the ontogenetic variation in the proteome of B. atrox and identified more P-III SVMPs in younger snakes than in adults, which could explain the higher hemorrhagic effects observed elsewhere [51].The difference in MHD observed between female and male pools in our work may be attributed to the different abundance of P-III SVMPs identified in the venom pools.[30] compared the geographic variation of B. atrox and reported a lower LD 50 than herein observed and suggested a correlation with the lower hemorrhagic activity.This is consistent with the results of the procoagulant and hemorrhagic activities, which are apparently related to the lethality of the venom [85,86].Another study relates a lack of hemorrhagic activity associated with a higher lethality in Daboia russelii [87].There was a marked difference between hemorrhagic and procoagulant activities between the venom of males and females, and these results may relate with the metabolic requirements of each sex.The metabolic rate of males and females is different, and it has been previously shown in viperids that females have a higher oxygen consumption, which is related to the animal's mass [88].Since B. atrox is a species displaying sexual dimorphism, in which females are usually larger than males, it is possible that females have a higher energy demand due to their larger size, in addition to the need of extra energy reserved for reproduction [89].
Regarding MHD, the variation may have been caused by the relative abundance of proteins with hemorrhagic activity, which is slightly lower in the female venom pool than in the male venom pool.This activity may be under the influence of other proteins and/or the synergistic effect of other compounds in the venom.

Immunorecognition by antibothropic serum
The antivenom produced at Butantan Institute is composed by antibodies raised in horses, using a mixture of B. jararaca (50%), B. jararacussu (12.5%),B. alternatus (12.5%),B. moojeni (12.5%) and B. neuwiedi complex (12.5%) venom.Although B. atrox is not included in the venom pool used to produce the antivenom, it seems to have a moderate reaction with the serum (Figure 4).
Overall, the antibothropic serum produced at Butantan Institute recognized all venoms similarly, especially the ones with higher and lower molecular weights (Figure 4).Curiously, the band between 20 and 25 kDa were not well recognized by the serum in all groups, although it's very strong in the gel (Figure 1B).Analyzing the MS (Table 1 and Additional file 3), it is concluded that this band probably represents a PI-SVMP.Other studies concerning B. atrox venom that also tested the immunerecognition using the antibothropic serum produced at Butantan Institute, showed that this reaction is not as strong as with other species' venom; and geographic variation seems to have great influence in the reactivity of the venoms to the antivenom [51,58,62,90].Moreover, Sousa and colleagues [30] found striking differences in the neutralization of in vivo activities of B. atrox venoms from different habitats, regardless of the similarity in the reaction observed by ELISA.

Conclusion
Several studies have shown that B. atrox venom may have variability in their biological activity and protein composition.This work extends the outlook regarding this variability, showing that female and male venoms of B. atrox siblings, under the same controlled environmental conditions, present subtle differences in their composition and activities.Moreover, it was observed individual variability in the characteristics of venoms, indicating that, in addition to aspects such as, geographical location, ontogeny, sex and diet, there are several unknown factors that result in the venom plasticity and physiological effects.kDa: kilodalton; LAAO: L-amino acid oxidase; LC-MS/MS: liquid chromatography-mass spectrometry/mass spectrometry; LD 50 : lethal dose 50%; MCD: minimum coagulant dose; MHD: minimum hemorrhagic dose; NOBA: 4-nitro-3-octanoyloxybenzoic acid; NT: nucleotidase; PDE: phosphodiesterase; PLA 2 : phospholipase A 2 ; PLB: phospholipase B; PVDF: polyvinylidene difluoride; RP-HPLC: reverse-phase high performance liquid chromatography; RP-UPLC: reverse-phase ultra performance liquid chromatography; SD: standard deviation; SVMP: snake venom metalloproteinase; SVSP: snake venom serine proteinase; TBS: Tris-buffered-saline; TCA: trichloroacetic acid; TFA: trifluoroacetic acid.

Figure 1 .
Figure 1.One-dimensional electrophoresis (1-DE) profile of B. atrox venoms under (A) non-reducing and (B) reducing conditions.Individual female (Ba1 to Ba5), male (Ba6 to Ba9) and respective pools were used and are indicated above the gel.
LD 50 of female venom pool of B. atrox (104.3 µg/animal; CI: 73.3−151.2µg/animal) was slightly lower than that of the male (118.4µg/animal; CI: 87.2−164.8µg/animal), but with no statistical difference.Although differences were observed in some activities, this is not reflected in the venom lethality.Saldarriaga et al. [47] found 81.4 µg/mice as LD 50 for adult B. atrox, a minor dose than found in this work.Also, Sousa et al.

Figure 4 .
Figure 4. Immune interaction between the proteins of B. atrox venoms and the antibothropic serum by western blotting.Individual female (Ba1 to Ba5), male (Ba6 to Ba9) and respective pools were used and are indicated above the gel.

Table 1 .
Identification of protein compounds found in female and male B. atrox venom pools, by LC-MS/MS.Proteins showing statistically different abundance (fold change ≥ 1.5 or ≤ 0.67; p < 0.05) are bolded.The last seven proteins listed were identified exclusively in the female venom pool.