Naja annulifera Snake: New insights into the venom components and pathogenesis of envenomation

Background Naja annulifera is a medically important venomous snake occurring in some of the countries in Sub-Saharan Africa. Accidental bites result in severe coagulation disturbances, systemic inflammation and heart damage, as reported in dogs, and death, by respiratory arrest, in humans. Despite the medical importance of N. annulifera, little is known about its venom composition and the pathogenesis of envenomation. In this paper, the toxic, inflammatory and immunogenic properties of N. annulifera venom were analyzed. Methodology/Principal findings Venom proteomic analysis identified 79 different proteins, including Three Finger Toxins, Cysteine Rich Secretory Proteins, Metalloproteinases, Phospholipases A2 (PLA2), Hyaluronidase, L-amino-acid oxidase, Cobra Venom Factor and Serine Proteinase. The presence of PLA2, hyaluronidase, fibrinogenolytic and anticoagulant activities was detected using functional assays. The venom was cytotoxic to human keratinocytes. In an experimental murine model of envenomation, it was found that the venom induced local changes, such as swelling, which was controlled by anti-inflammatory drugs. Moreover, the venom caused death, which was preceded by systemic inflammation and pulmonary hemorrhage. The venom was shown to be immunogenic, inducing a strong humoral immune response, with the production of antibodies able to recognize venom components with high molecular weight and to neutralize its lethal activity. Conclusions/Significance The results obtained in this study demonstrate that N. annulifera venom contains toxins able to induce local and systemic inflammation, which can contribute to lung damage and death. Moreover, the venom is immunogenic, an important feature that must be considered during the production of a therapeutic anti-N. annulifera antivenom.


Introduction
Envenoming from snakebites is a public health problem in rural areas of the tropical and subtropical countries in Africa, Latin America, Asia and Oceania [1,2]. This medical condition kills more than 95,000 people per year and leads over 300,000 victims to live with permanent sequelae [3].
It is also estimated that there are approximately 3,700 species of snakes worldwide [4]. Of these, approximately 15% are venomous and have caused serious accidents involving humans and other animals [5]. Venomous snakes belong to the Colubroidea superfamily, which is composed of several families, such as Colubridae, Viperidae, Lamprophiidae and Elapidae [6,7]. The Elapidae family consists of 61 genera and includes 365 species [8], which are distributed in the tropical and temperate regions of Africa, America, Asia and Australia. These snakes can live in terrestrial or aquatic environments and present variable diet, including small vertebrates, such as birds, rodents, reptiles and fishes or invertebrates [9,10].

Venoms
N. annulifera venom (South Africa specimens) was purchased from Latoxan Natural Active Ingredients (Valence, France). The lyophilized venom was reconstituted in sterile saline solution at 5 mg/mL. The protein content was assessed with the BCA Protein Assay Kit according to the manufacturer's recommendations. The samples were aliquoted and stored at -80˚C until use. Bothrops jararaca, Crotalus durissus terrificus and Tityus serrulatus venoms were supplied by Butantan Institute, SP, Brazil, and their use was approved by the Brazilian Institute of Environment and Renewable Resources (IBAMA), an enforcement agency of the Brazilian Ministry of the Environment (protocol number: 010035/2015-0), and by SisGen (Sistema Nacional do Patrimônio Genético e do Conhecimento Tradicional Associado (protocol numbers AD50761 and AEE9AEA).

Ethics statement
High III (H III ) female mice weighing 18-22 g were obtained from the Immunogenetics laboratory, while Balb/c male mice weighing 18-22 g were obtained from the Center for Animal Breeding, both from Butantan Institute. All procedures involving animals were in accordance with the ethical principles for animal research adopted by the Brazilian Society of Animal Science and the National Brazilian Legislation n˚. 11.794/08. The protocols used in the present study were approved by the Institutional Animal Care and Use Committee of the Butantan Institute (protocols approved n˚01092/13 and 01262/14).
Experiments using samples obtained from humans were previously approved by the Human Research Ethics Committee of the Municipal Health Secretary of São Paulo. Human blood samples were obtained from healthy donors who knew of the purposes of this study and signed the corresponding informed consent form (protocol approved n˚974.312).
Mass spectrometric protein identification. For the proteomic analysis, N. annulifera venom was submitted to trypsin digestion as described by Kinter and Sherman [47]. Briefly, 200 μg venom (three replicates) were denatured with 6 M urea in 100 mM Tris-HCl, pH 8, and reduced with dithiothreitol (DTT) 10 mM for 1 h at room temperature. Cysteine carbamidomethylation was performed by incubation with iodoacetamide 40 mM for 1 h at room temperature, followed by addition of DTT 40 mM to consume excess of alkylating agent. Protein samples were diluted with water until the concentration of urea decreased to about 0.7 M and trypsin was added using an enzyme: substrate ratio of 1:50 (w/w) at 37˚C for 18 h. The reaction was stopped with 20% TFA until pH 3 and desalting was carried out by loading individual samples to p-1000 StageTip SDB-XC [48]. After drying using a vacuum centrifuge, peptide mixtures were dissolved in 100 μL of 0.1% formic acid (solution A) and peptide concentration was estimated. For that, a peptide standard curve was prepared using a mixture of trypsindigested bovine serum albumin (MassPREP BSA Digestion Standard; Waters) ranging from 0.2 to 1.2 μg/μL (in solution A). The absorbance was measured at 214 nm using a NanoDrop spectrophotometer (Thermo Scientific), and peptide sample concentration was determined using the standard curve. For analysis, each peptide sample (5 μg) was automatically injected using an EASY Nano LCII system (Thermo Scientific) into a 5 cm of 10 μm Jupiter C-18 trap column (100 μm I.D. × 360 μm O.D.) coupled to an LTQ-Orbitrap Velos mass spectrometer (Thermo Scientific). Chromatographic separation of tryptic peptides was performed on a 15-cm long column (75 μm I.D. x 360 μm O.D.) packed in-house with 3 μm ReproSil-Pur C-18 beads (Dr. Maisch, Germany). Peptides were eluted with a linear gradient of acetonitrile in 0.1% formic acid (solution B) at 200 nL/min: the linear step used for the elution was 5-35% in 75 min. Spray voltage was set at 2.2 kV and the mass spectrometer was operated in data dependent mode, in which one full MS scan was acquired in the m/z range of 300-1,600 followed by MS/MS acquisition using collision induced dissociation of the ten most intense ions from the MS scan. MS spectra were acquired in the Orbitrap analyzer at 30,000 resolution (at 400 m/z) whereas the MS/MS scans were acquired in the linear ion trap. The minimum signal threshold to trigger fragmentation event, isolation window, activation time and normalized collision energy were set to, respectively, 1,000 cps, 2 m/z, 10 ms and 35%. A dynamic peak exclusion (list size of 500) was applied to avoid the same m/z of being selected for the next 20 seconds. Two independent LC-MS/MS runs were performed for each sample.
Database search. Protein identification was performed using the MaxQuant software suite (v 1.5.3.12) [49] using the theoretical trypsin fragments produced from the database restricted to the taxonomy "Serpentes" (Uniprot release September 14, 2016; 63,114 sequences). The database search considered the precursor mass tolerance of 20 ppm, in the first search, and the nonlinear mass recalibration was made in the masses of first pass identifications. The second search considered the mass tolerance of 4.5 ppm and tolerance for fragments was set to 0.5 Da. The following search parameters were used: iodoacetamide derivatives of cysteine were considered as fixed modification, oxidation of methionine was specified as variable modification, and up to two missed cleavages by trypsin were accepted. Assignments for peptides and proteins were accepted at a false discovery rate < 1%. For the identification, only proteins identified in at least two experimental replicates, with at least 1 unique peptide and posterior error probability � 0.01 were accepted. "Match between runs" feature was enabled. For each protein group in the MaxQuant's 'proteinGroups.txt' file, the first protein entry was selected as representative. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org), via the PRIDE partner repository [50], with the dataset identifier: PXD010962 (Username: reviewer27828@ebi.ac.uk; Password: BNxFCLRG).

Toxic-Enzymatic properties of the venom
Gelatinolytic activity. The gelatinolytic activity of the N. annulifera snake venom was evaluated with zymography using gelatin as the substrate [51]. Samples of the venom (30 μg), prepared under non-reducing conditions, were separated via 10% SDS-PAGE with 1 mg/ml gelatin. The gels were washed for 2 hours at room temperature in 2.5% Triton X-100 solution and incubated overnight at 37˚C in substrate buffer (50 mM Tris-HCl, 200 mM NaCl, 10 mM CaCl 2 , 0.05% Brij-35, p.H. 8.3). After this period, the gels were stained with Coomassie Brilliant Blue solution (40% methanol, 10% acetic acid and 0.1% Coomassie Brilliant Blue).
Fibrinogen cleavage. Human fibrinogen samples (30 μg) were incubated with 5 μg of N. annulifera venom or saline for 1 hour at 37˚C under constant agitation. In parallel, venom samples were pre-incubated with 1.10 Phe (10 mM) or PMSF (10 mM) at room temperature for 30 minutes. After incubation, the mixtures were submitted to SDS-PAGE using a 8-16% gradient gel under reducing conditions and stained with Coomassie Brilliant Blue solution. The positive control of the reaction was human Thrombin (1U).
Hyaluronidase activity. Hyaluronidase activity was analyzed according to the methodology described by Purkrittayakamee et al. [52] with slight modifications. Briefly, samples of N. annulifera venom (20 μg) were incubated with a mixture that contained hyaluronic acid (1 mg/mL) (100 μL) and acetate buffer (200 mM sodium acetate and 0.15 M of NaCl, p.H. 6.0) (400 μL) at 37˚C for 15 minutes. After incubation, the reactions were stopped with 1 mL of CTAB (2.5% CTAB and 2% NaOH). The absorbances were measured at λ405 nm in a spectrophotometer (Multiskan EX, Labsystems, Finland) against a blank that contained hyaluronic acid, acetate buffer and CTAB. The results are expressed as units of turbidity reduction (UTR) per mg of venom. T. serrulatus scorpion venom (20 μg) was used as a positive control. PLA 2 activity. PLA 2 activity was evaluated using the EnzChek TM Phospholipase A2 Assay Kit according to the manufacturer's recommendations. Briefly, samples of N. annulifera venom (0.5 μg) was incubated with a phospholipid mix that contained 10 mM Dioleoylphosphatidylcholine and 10 mM Dioleoylphosphatidylglycerol in 96-well microtiter plates. Increased fluorescence was evaluated using a spectrometer FLUOstar Omega (BMG Labtech, Ortenberg, Germany) at λ EM 460 and λ EX 515 nm and 37˚C for 10 minutes. Specific activity was expressed as Units of Fluorescence (UF) per minute per microgram of venom. C. d. terrificus (0.5 μg) venom was used as positive control.
Cytotoxic activity. The HaCat human keratinocyte cell lineage was cultured in DMEM supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin at 37˚C and 5% CO 2 . The action of N. annulifera venom on the viability of human keratinocytes was evaluated using the MTT (3-(4,5-dimethylthiazol-2-yl) -2,5-diphenyltetrazolium bromide) method [53] with minor modifications. Keratinocytes (5×10 4 cells/well) were grown in 96-well plates that contained supplemented DMEM medium and were then maintained for 24 hours in DMEM medium without SFB. After this period, cells were incubated for 72 hours in increasing venom concentrations at 37˚C and 5% CO 2 . The supernatant from the wells was aspirated and 60 μl of the MTT solution (0.83 μg/μL) in incomplete DMEM medium was added to the wells. The plates were then incubated for 30 minutes at 37˚C and 5% CO 2 . Thereafter, the supernatant was aspirated, 100 μL/well of DMSO was added, and the plates were allowed to stand at room temperature for five minutes. Additionally, the action of N. annulifera venom on the viability of human keratinocytes was evaluated by Lactate Dehydrogenase (LDH) release assay. Supernatants of keratinocyte cultures, treated for 72 hours with increasing venom concentrations, were centrifuged at 405× g for 10 min, and tested for the presence of LDH, using the CytoTox 96 Non-Radioactive Cytotoxicity Assay Kit, according to the manufacturer's instructions.

The impact of venom on hemostatic parameters
Blood samples from healthy donors were placed into tubes containing sodium citrate (3.2%) and centrifuged at 260× g at room temperature to obtain platelet-poor plasma (PPP). The PPP samples were aliquoted and stored at -20˚C until their use.
Activated Partial Thromboplastin Time (APTT). PPP samples (50 μL) were treated with saline or increasing concentrations of venom and incubated for 3 minutes at 37˚C with 50 μL of Cephalin. PPP was then recalcified with CaCl 2 (0.025 M), and the coagulation time was measured over 240 seconds in a Stago Start System (Asnières-sur-Seine, France). The results are expressed in seconds and R-time, which was obtained from the experimental sample/control sample. Disorders in coagulation time were considered significant when the Rtime was over 1.3 [54].
Prothrombin Time (PT). PPP samples (50 μL) were treated with saline or increasing concentrations of venom and incubated for 1 minute at 37˚C. PPP was then incubated with CaCl 2 (0.025 M) that contained thromboplastin (100 μL), and the coagulation time was followed for 60 seconds using the Stago Start System (Asnières-sur-Seine, France). The results are expressed in seconds and R-time, which was obtained from the experimental sample/control sample. Abnormalities in coagulation time were considered significant when the R-time was over 1.3 [54].
Local reactions: Edematogenic activity and pharmacological modulation. The venom edematogenic activity was assessed via the method previously published by Yamakawa et al. [57] with some modifications. Balb/c mice (n = 6) were inoculated with 10 μg of N. annulifera venom at a volume of 50 μL (sterile saline) in the subcutaneous (s.c) tissue of the plantar region of the animal's left hindpaw. The contralateral hindpaw (control) was inoculated with 50 μL of sterile saline. To assess edematogenic activity, the thickness of the hindpaw was evaluated with a caliper rule (Mitutoyo, Suzano-SP, Brazil; sensibility of 0.01 mm) at different time points before (T0) and over the 24 h, following either venom or saline inoculation (Te). Increases in paw volume are expressed in percentage form (%), calculated with the following formula: (Te-T0)/T0 � 100.
Systemic reactions. To analyze the possible systemic reactions induced by N. annulifera venom, two experimental groups were established. The first measured systemic alterations over time. LD 50 assays were performed to define the sublethal dosing level that was able to induce the deleterious effects of the venom, including an increase in inflammatory parameters, but not kill the animals. LD 50 values that varied between 50 and 75% were tested, and the selected dose for the inflammatory assays was 60% of LD 50 . Balb/c mice (n = 6) were then inoculated with 60% of LD 50 or sterile saline i.p. and euthanized at different time points to obtain blood and collect the organs.
In another experimental set, it was assessed whether death promoted by venom was preceded by systemic reactions. Animal groups (n = 6) were inoculated with saline or 2 LD 50 of N. annulifera venom i.p. Immediately after animals' death, their blood and organs were collected. Blood samples were obtained by cardiac puncture and were immediately incubated with EDTA (2.5 mg/mL). Aliquots of these samples were used to measure systemic leukocyte changes, while other blood samples were centrifuged at 2800× g at 4˚C for 10 minutes to isolate plasma. Plasma samples were stored at -80˚C and used to measure inflammatory mediators.
Leukocyte alterations. To analyze changes in the total number of leukocytes, blood samples were diluted 1:10 in Turk's solution (0.1% crystal violet dye in 2% acetic acid), and the cells were counted in a Neubauer chamber. Blood samples were also submitted for blood smear and stained with a Fast Panoptic stain kit. Differential cell counts were performed by analyzing 100 cells that were identified as lymphocytes, neutrophils or monocytes based on morphologic criteria.
Detection of inflammatory mediators. IL-6, TNF-α, MCP-1, IL-12, IFN-γ and IL-10 detection was performed using the BD Cytometric Bead Array (CBA) Mouse Inflammation Kit (BD Bioscience). IL-1β and IL-17 quantification was performed using the Mouse IL-1β ELISA (BD Bioscience) and Mouse IL-17 Duo3.2.43 ELISA (R&D Systems) kits, respectively. All assays were performed according to the manufacturer's instructions. The results are expressed as pg/mL. Histopathologic analysis. To analyze systemic histopathologic changes, Balb/c mice (n = 6) were inoculated i.p. with a sublethal dose or 2DL 50 of N. annulifera venom. After euthanasia or death, animals were exsanguinated, and their brains, lungs, hearts, kidneys, spleens and livers were collected. To analyze local histopathological changes, Balb/c mice (n = 3) were injected subcutaneously with 10 μg of N. annulifera venom at a volume of 50 μL (sterile saline) in the plantar region of their left hindpaws. The contralateral hindpaws (control) were inoculated with 50 μL of sterile saline. Animals were euthanized at different time points (0, 20, 60, 240 and 1440 minutes) and had their hindpaws removed. The hindpaws and organs were fixed in a 10% formaldehyde solution for 24 hours, submitted for routine histology and stained with hematoxylin and eosin (HE). All samples were analyzed with a light microscope and were examined for the presence of necrosis, edema, inflammatory infiltrates or hemorrhagic foci.

Experimental antivenom production
Immunogenicity. H III mice (n = 12) were immunized subcutaneously with 10 μg of N. annulifera venom diluted 1:25 on Al(OH) 3 to achieve a final volume of 200 μL. These animals received three booster doses (5 μg) at 20-day intervals. Control animals were inoculated with sterile saline. Bleeding was performed 10 days after each venom inoculation. Blood was allowed to clot at room temperature for 15 minutes and was then left at 4˚C for 6 hours. After centrifugation at 252× g and 4˚C, the sera were collected and immediately frozen at -20˚C.
Sera antibody titers. To evaluate the sera antibody titers of immunized animals, ELISA plates (Costar) were coated with 100 μL of N. annulifera venom (10 μg/mL) and incubated overnight at 4˚C. The plates were then blocked with 5% BSA in PBS for 2 hours at 37˚C and incubated with crescent dilutions of non-immune or experimental sera samples for 1 hour at room temperature. After incubation, the plates were washed with PBS-Tween 20 0.05%, and incubated with the IgG-HRPO (1:5000 diluted) secondary antibody in 1% BSA/PBS for 1 hour at 37˚C. The plates were then washed, and the reactions were developed with OPD substrate according to the manufacturer's instructions. The optical densities were obtained using spectrophotometry (Multiskan EX, Labsystems, Finland) at λ492 nm.
Recognition of venom components using experimental antivenom. Samples of N. annulifera venom (15 μg) were separated using 8-16% gradient SDS-PAGE [43] and blotted on nitrocellulose membranes [45]. The membranes were blocked for 2 hours with 5% BSA in PBS, washed three times with PBS, incubated with normal or experimental sera pools, and diluted 1:5000 in 0.01% PBS/BSA for 1 hour at room temperature. After incubation, the membranes were washed with PBS-0.05% Tween 20, incubated with IgG-AP secondary antibodies, and diluted 1:5000 in 1% BSA/PBS for 1 hour at room temperature. After incubation, the membranes were washed three times in PBS-0.05% Tween 20. Antigenic reactivity was measured by the addition of NBT and BCIP substrates according to the manufacturer's recommendations.
Serum neutralization of the venom's lethal activity. To determine the neutralizing potential of the experimental antivenom, samples of the venom, equivalent to 2LD 50 (188.28 μg), were incubated with saline or crescent dilutions of the experimental serum for 30 minutes at 37˚C. These mixtures were then i.p. administered to Balb/c mice (n = 6) (200 μL). Animals were observed for 72 hours and the death/survival ratio recorded. The neutralizing potential was calculated by Probit analysis and the results expressed as potency (mg of venom neutralized by mL of experimental serum) [56].

Statistical analysis
Statistical analysis was performed using one-way ANOVA followed by Tukey's Multiple comparison test, two-way ANOVA followed by a Bonferroni multiple comparison or t-tests. All statistical analyses were performed using Graphpad Prism 5 software (La Jolla, California, USA). Differences were considered significant when p � 0.05.

N. annulifera venom composition
The analysis of the electrophoretic profile of N. annulifera venom using a 8-16% SDS-polyacrylamide gel, under non-reducing conditions, revealed protein bands with relative molecular masses between~10 kDa and~190 kDa. Under reducing conditions, the venom displayed a more simple profile, with main protein bands ranging from~6 kDa to~100 kDa (Fig 1A), indicating that the venom contains protein in oligomeric state. To better characterize venom proteins, nitrocellulose membranes to which venom samples had been transferred were incubated with the ConA and WGA lectins, which recognize mannose [64] and N-acetylglucosamine [65] residues, respectively. Fig 1B shows three protein bands (~70,~86 and~103 kDa) that interacted with WGA indicating the presence of N-acetylglucosamine. Moreover, three bands that interacted with ConA (~63,~81 and~170 kDa) indicated that these proteins contain mannose residues in their carbohydrate moieties (Fig 1C).
The venom proteomic analysis using in-solution trypsin digestion and LC-MS/MS resulted in the identification of 79 proteins, including 24 types of venom components with or without enzymatic function, and six proteins of unknown function. Despite the variety of proteins identified in the venom, it is dominated by the presence of 3FTx (27 proteins) ( Table 1; S1 and S2 Tables; S1 Spectra). Other protein classes identified in the venom include cysteine-rich secretory protein (CRISP), Kunitz-type protease inhibitor, snake venom metalloproteinase (SVMP), snake venom serine proteinase (SVSP), phospholipase A2 (PLA 2 ), hyaluronidase (glycosyl hydrolase 56 family) and Cobra Venom Factor (venom complement C3 homolog).

Toxic-Enzymatic properties of N. annulifera venom
N. annulifera venom showed no proteolytic activity on zymography using gelatin as substrate. Under the same experimental conditions, proteins from the positive control of B. jararaca venom showed gelatinolytic activity, as demonstrated by clear regions in the gel (bands with molecular mass above 60 kDa) (Fig 2A). However, when proteolytic activity on fibrinogen was assessed, it was observed that the venom contained proteinases that were able to cleave this protein at the alpha chain, generating a fragment with a molecular mass of~40 kDa (Fig 2B). In addition, when inhibitors were added to the reactions, cleavage was inhibited by 1, 10 Phe and PMSF, demonstrating the contributions of SVMP and SVSP to this hydrolysis (Fig 2B).  The hyaluronidase activity of N. annulifera snake venom was assessed via a turbidimetric assay, and the results are expressed in UTR. It was demonstrated that N. annulifera venom has a low hyaluronidase activity (UTR = 14.4) compared to the positive control T. serrulatus scorpion venom (UTR = 52.8) (Fig 2C). PLA 2 activity was measured with a fluorimetric assay, and the results obtained are expressed as specific activity (SA: UF/min/μg venom). Fig 2D shows that N. annulifera snake venom has significantly lower phospholipase activity (SA: 566.2) than the C. d. terrificus venom positive controls (SA: 3061.3).  Cytotoxic properties of the venom were evaluated by cell viability and LHD release assays. N. annulifera venom was able to reduce human keratinocyte viability in a dose-dependent manner but was not able to promote LDH release by the cells (Fig 2E).

The impact of N. annulifera on hemostasis
N. annulifera venom induced disturbances in the Activated Partial Thromboplastin Time in a dose dependent manner ( Table 2). Platelet-poor plasma (PPP) samples were incoagulable when the highest venom concentrations were used (12.5 to 50 μg), while the coagulation time (seconds) was significantly prolonged at lower concentrations (2.5 to 6.25 μg) (p� 0.05). Moreover, the R-time shows that the disturbances promoted by N. annulifera venom were severe, as they were much higher than normal. Prothrombin Time assays demonstrated that the venom led to a significant prolongation in the coagulation time of PPP at the highest venom concentrations, i.e., 25 and 50 μg. Besides that, both venom concentrations were able to cause a significant alteration in the R-time ( Table 2).

Toxicity of N. annulifera venom in the murine model
Local reactions. To evaluate the edematogenic activity of N. annulifera venom, Balb/c mice were inoculated with 10 μg of venom into their left hind footpads, and their paws were measured at different time points. The venom was able to induce a significant rapid onset of edema, with its maximum peak (110.13% of increase on paw volume) reached 20 minutes after inoculation. It returned to basal level 24 hours after venom inoculation (Fig 3A).
In addition to the evaluation of edematogenic activity, the contribution of mast cell and lipid mediators to edema development was also assessed. It was found that mast cell degranulation is important for the earliest development (10-30 minutes) of edema as treatment with Cromolyn, a mast cell degranulation inhibitor, was able to reduce swelling (S1A Fig). Dexamethasone, a synthetic corticosteroid compound with a potent anti-inflammatory activity, had a strong effect on the paw edema induced by N. annulifera venom. It was able to decrease hindpaw thickness throughout the duration of edema, mainly during the early stages (10-30 minutes), suggesting that lipid mediators participate in the development of edema (S1B Fig). Indomethacin, a non-selective COX inhibitor, was also able to reduce the degree of edema (10 analysis was performed using the t-test ( ��� p� 0.05).
[D] PLA 2 activity: Samples of N. annulifera venom (0.5 μg) were incubated at 37˚C with a phospholipid mix that contained 10 mM phosphatidylcholine and 10 mM phosphatidylglycerol. Increased fluorescence was measured for 10 minutes. As a positive control, C. d. terrificus venom (0.5 μg) was used. The results are representative of three separate experiments and expressed as specific activity (UF per μg of venom per minute) ± SD. Statistical analysis was performed using t-test ( ��� p� 0.05).
[E] Cytotoxic activity: The HaCat human keratinocyte cell lineage was cultured in DMEM medium and incubated during 72 hours with increasing amounts of N. annulifera venom. The effect on cell viability was evaluated using the MTT method and by measuring LDH release from human keratinocytes exposed to the venom, using the CytoTox 96 Non-Radioactive Cytotoxicity Assay Kit. Statistical analysis was performed using One Way ANOVA ( ��� p� 0.05) ± SD.
https://doi.org/10.1371/journal.pntd.0007017.g002 to 120 minutes), suggesting that prostaglandins and thromboxanes may contribute to the local reaction to the venom. However, treatment with Indomethacin was not able to control swelling during the late phase (180 to 1440 minutes) (S1C Fig). A similar effect was observed in animals treated with a FLAP inhibitor (S1D Fig). The PAFR blocker WEB-2086 demonstrated inhibitory action only in the earliest phases of the edema (20-30 minutes) (S1E Fig). Although   Fig 3. Local alterations induced by N. annulifera venom. [A] Edema: samples of venom (10 μg/50 μL) or sterile saline were injected into the left or right hind footpad, respectively, of Balb/c mice (n = 6). Paw edema was assessed by measuring the paw thickness using a caliper rule before (T0) and after inoculation with venom or saline (Te). The increased paw volume was expressed as percentage (%) and was calculated with the following formula: (Te-T0)/ T0 � 100. Statistical analysis was performed using two-way ANOVA, followed by the Bonferroni multiple comparison test ( ��� p� 0.05) ± SD.
[B] The effect of inhibitors on the paw edema induced by N. annulifera venom: groups of Balb/ c mice (n = 6) were treated with different inhibitors before induction of edema with venom: Cromolyn, a mast cell degranulation inhibitor (10 mg/kg, three days consecutively before the edema, i.p. route); dexamethasone (Dexa), a cPLA 2 inhibitor (2 mg/kg, 2 hours before the edema, i.p. route); indomethacin (Indo), a COX isoform inhibitor (10 mg/kg, 30 minutes before the edema, i.p. route); MK886, a 5-lipoxygenase-activating protein inhibitor (5 mg/kg, 30 minutes before the edema, i.p. route); and WEB2086, a PAFR antagonist (5 mg/kg, 1 hour before the edema, s.c. route). After treatment, edema was induced, and the paws were measured with a caliper rule at several time points. The increased paw volume was expressed as a percentage. Statistical analysis was performed using two-way ANOVA, followed by the Bonferroni multiple comparison test ( ��� p� 0.05) ± SD. all inhibitors showed different modulation profiles, they were able to significantly decrease the edema at its peak [ Fig 3B]. Histopathologic analyses showed that the venom promotes several changes, ranging from mild to intense, at all time periods [ Fig 3C]. After inoculation with the venom, it was observed that it promotes an acute inflammatory response at the inoculation site, with a disorganized dermal collagen matrix, subcutaneous edema, vascular congestion [Panel 2], mild erythrocyte extravasation and neutrophil migration to the dermis and muscle [Panels 3,4,5]. Myonecrosis was also observed [Panels 4, 5] along with polymorphonuclear cell infiltration. These alterations persisted for several hours. However, after 24 hours of inoculation, a decrease in the subcutaneous edema was observed, along with the absence of vascular congestion and few remaining neutrophils in the tissue. By contrast, an increased number of mononuclear cells and reorganization of the dermal collagen matrix was observed [Panel 6].

Lethal dose
LD 50 was determined 72 hours after venom administration. Prior to their death, animals presented with several clinical findings of envenomation, including apathy, bending at the column, a rough hair coat, dyspnea and paralysis of their hind limbs. LD 50 was calculated with a probit analysis with a 95% confidence interval and was found to be 94.14 μg (68. 78-115.74).

Systemic reactions induced by N. annulifera venom
The systemic reactions promoted by the venom were evaluated using two different experimental protocols. By injecting 60% (56.48 μg) (Sublethal dose) of LD 50 of N. annulifera venom, it was observed that animals presented several clinical findings of envenomation, including apathy, bending at the column, a rough hair coat, dyspnea and difficulty walking. This dose was also able to cause alterations in some systemic parameters, such as a decrease of circulating lymphocytes [ Fig 4A] and an increased number of circulating neutrophils [ Fig 4B]. Nonetheless, this dose was not able to lead to changes in the total number of leukocytes and histological organs damage. However, a sublethal dose promoted an increase in the plasma levels of MCP-1 [ Fig 4C] and IL-6 [ Fig 4D]. These alterations persisted for several hours. However, 24 hours after experimental envenomation, all of these parameters returned to their normal values.
Injection of 2LD 50 led to several clinical findings of envenomation, such as apathy, bending at the column, a rough hair coat, dyspnea, hind limb paralysis and death. In addition to these clinical findings, histopathologic alterations were observed in the lungs, among these, moderate vascular congestion and multiple hemorrhagic foci [ Fig 4I]. Furthermore, the animals demonstrated leukocytosis [ Fig 4E] characterized by neutrophilia and monocytosis [Fig 4F], as well as increased plasma levels of MCP-1 [ Fig 4G] and IL-6 [ Fig 4H].

N. annulifera venom immunogenicity and antivenom production
The immunogenicity of N. annulifera snake venom was evaluated in H III mice immunized with 10 μg of venom. The antibody titers were determined by ELISA. Fig 5A presents the antibody response over time. N. annulifera venom is highly immunogenic, promoting the production of high antibody titers, already after the second immunization dose. The recognition profile of the venom proteins by the experimental N. annulifera mouse antiserum was determined by Western Blot. Fig 5B shows that the experimental serum was able to recognize venom components, mainly proteins with Mr above 50 kDa. Moreover, as determined by Probit analysis, it was shown that this antivenom was able to neutralize the N. annulifera venom lethal activity with a high potency, i.e., 1 ml of antivenom was able to neutralize 4.5 mg of venom.

Discussion
In this study, we characterized some of the biochemical, toxic, immunogenic and physiopathologic properties of the venom from N. annulifera, which is a medically important snake related to accidental bites in the countries of Sub-Saharan Africa. The results show that N. annulifera venom contains several toxic components able to induce systemic inflammation, which may contribute to the pathology observed in envenomed individuals. Moreover, the venom is immunogenic, an important feature that must be considered during the production of a therapeutic anti-N. annulifera antivenom.
Electrophoretic analysis of N. annulifera venom showed that it contains several components, including low molecular mass proteins, which suggested the presence of neurotoxins [18], among these, 3FTx [14,15] and PLA 2 [16]. The presence of these components in the venom was confirmed by trypsin digestion of proteins and LC-MS/MS analysis. These components may be responsible for some of the clinical findings observed during envenomation, such as heart damage and systemic inflammation in dogs [35], as well as respiratory arrest in humans [23]. The data from this study corroborate the results of other studies, which showed the presence of these components in N. annulifera venom [37,38,39,40,41,42]. Envenomed  individuals from South Africa showed local dermonecrotic injury after a bite, which could be caused by the high content of cytotoxins, as shown in our LC-MS/MS analysis. In accordance with data from Panagides and colleagues [66], here we also showed that N. annulifera venom could promote decrease in human epidermal cells viability, as evaluated by the MTT method. However, it was not possible to detect release of LDH by these cells, which possibly indicate that the keratinocyte membranes were not damaged and that the cell death promoted by N. annulifera could be due to apoptosis [67] and not necrosis. Alternatively, it is possible to consider that N. annulifera venom contains components able to reduce mitochondrial activity, since MTT method evaluates cell viability as enzymatic conversion of the tetrazolium compound to water insoluble formazan crystals by dehydrogenases occurring in the mitochondria of living cells.
Western Blot lectin analysis demonstrated the presence of mannose and N-acetylglucosamine residues in N. annulifera venom proteins. These carbohydrate residues were found in the venom proteins of different genera of snakes [46,68,69,70], including in important toxic components, such as SVSP and SVMP [69,70]. Although we have not identified the families of these glycosylated components, it is possible that some of them are linked to SVMPs or SVSPs, as they have predicted molecular masses similar to some of the previously described glycosylated proteolytic enzymes from viperid venoms [69,70]. Moreover, SVMPs and SVSPs were detected in N. annulifera venom by LC-MS/MS. Proteins containing these carbohydrate residues are also present in many pathogens, such as bacteria and fungi, and can be recognized by different immune cells and molecules, triggering inflammatory and immune responses [71,72]. This recognition could contribute to the clinical manifestations observed during envenomation, such as the systemic inflammation observed in dogs [35].
Functional biochemical assays were performed to confirm the presence of some of the toxic-enzymatic components found in the LC-MS/MS analysis. The presence of proteinases in animal venoms can contribute to different clinical manifestations during envenomation, such as inflammation, tissue damage, disturbances in coagulation and bleeding [73]. In elapid venoms, proteolytic activity is usually low or nonexistent [26,74,75]. In contrast to the data shown by Phillips et al. [76], under the experimental conditions used in this paper, N. annulifera venom did not show any proteolytic activity on either gelatin. However, although detected at low abundance in the proteomic analysis, the venom contains SVMP and SVSP able to cleave the fibrinogen alpha chain, which suggests that these proteinases can contribute to the hemostatic alterations as observed in dogs envenomated by N. annulifera [36].
Hyaluronic acid is cellular cement that, together with other components of the extracellular matrix, forms a protective gel that prevents the entry of foreign agents. A variety of animal venoms contain hyaluronidases, which cleave hyaluronic acid molecules to facilitate access of the venom from the tissue to the bloodstream. This enzyme is also called "spreading factor" [77], and its action can promote local and systemic inflammation by increasing tissue permeability [78]. Moreover, the products derived from hyaluronic acid cleavage, which can be recognized by immune receptors, such as TLR2 and 4, trigger the production of inflammatory mediators, such as cytokines and chemokines [79]. N. annulifera venom showed evidence of hyaluronidase activity, but it was relatively low compared with other elapid venoms from the Naja [74] and Micrurus [16] genera, and accordingly, only one hyaluronidase was identified by mass spectrometry in the venom.
PLA 2 belongs to a superfamily of lipolytic enzymes that catalyze specific hydrolysis of the ester linkage at the sn-2 position of glycerophospholipids, generating arachidonic acid and lysophospholipids. PLA 2 -like proteins found in snake venoms may be devoid of catalytic activity, although it may exhibit myotoxic or neurotoxic activities [80,81]. Moreover, these enzymes may present with other toxic properties during envenomation, such as the ability to cause cytotoxicity and inflammation [82,83]. As predicted by LC-MS/MS, the presence of PLA 2 activity in N. annulifera venom was also observed in enzymatic assays. Nonetheless, N. annulifera venom exhibits a very lower content of PLA2 /enzymatic activity when compared to other Naja venoms [84,85,86,87,88], suggesting a strong interspecific variation associated to this particular toxin.
As observed in dogs, we showed here that N. annulifera venom promoted hemostatic disturbances in human plasma, making it incoagulable. These disturbances in the hemostatic system can be attributed to the fibrinogenolytic proteinases detected in the venom, or to PLA 2 found in the LC-MS/MS analysis, since these components can promote plasma incoagulability via direct binding to FXa, thereby preventing thrombin generation [89,90,91,92]. This phenomenon, promoted by several species of Naja venom, has been observed in different clinical and experimental studies [93,94]. It is therefore very important to evaluate the mechanisms involved in these alterations and their consequence, since the current literature on the topic is scarce. In addition, these alterations may be a therapeutic target for Naja envenomation.
In our in vivo experimental model, the venom was able to induce swelling and several histopathologic changes in the hind paws of mice, that decreased only after 24 hours. Among these tissue alterations, myonecrosis associated with inflammation was observed, an event that is commonly found in experimental models of Elapidae envenomation, which is attributed to cytotoxins and PLA 2 [95,96].
The inflammatory events promoted by the venom may be attributed to hyaluronidase, PLA 2 , glycosylated proteins, SVSP and SVMP since they can promote different inflammatory events, which include complement activation, mast cell degranulation, release of eicosanoids and cytokines and leukocyte homing [16,26,27,58,59,60,61,83]. Knowing that N. annulifera venom promotes inflammation and pain in humans and dogs [31,32,33,34,35,36], which were also observed in our experimental model, pharmacologic studies were performed to analyze the role of some inflammatory mediators in the edema process. Pre-treating mice with different compounds that were able to modulate different steps of the inflammatory process, including, mast cell degranulation (Cromolyn), lipid mediator production (Dexamethasone, Indomethacin and MK-886) and action (WEB-2086) significantly decreased the edema promoted by N. annulifera venom. All of the compounds showed a similar pattern of inhibiting peak edema. However, cPLA 2 , COX isoforms and FLAP inhibitors controlled the edema for a long time, suggesting a stronger contribution of eicosanoids, such as prostaglandins, thromboxanes and leukotrienes, to this process. Moreover, it is possible that these lipid mediators may contribute to the pain observed in humans after a bite, making them good therapeutic targets for the local reactions promoted by N. annulifera venom.
The LD 50 of N. annulifera, established here via i.p. route in Balb/c mice, was 94.14 μg. Ramos-Cerrillo and collaborators [56] showed that the N. annulifera venom LD 50 when administered intravenously was 53.9 μg. To evaluate whether N. annulifera venom was able to promote systemic changes, a sublethal dose of the venom was established, and different inflammatory parameters were evaluated. As in dogs envenomed by N. annulifera [34], it was observed that a sublethal dose of the venom promotes acute systemic inflammation, which was characterized by neutrophilia and increased levels of IL-6 and MCP-1 in the plasma. However, unlike dogs envenomed by N. annulifera the sublethal dose was not able to cause organ injury, as observed by Langhorn et al. [35] in dogs.
By administering a superdose (2LD 50 ) of venom to mice, it was possible to observe systemic inflammation, which was characterized by an increase in the plasma levels of IL-6 and MCP-1. However, in contrast with the sublethal dose, the 2DL 50 dose caused leukocytosis, which was characterized by neutrophilia and monocytosis. In addition, dead animals showed multifocal hemorrhaging in their lungs.
These data suggest that the systemic inflammatory process induced by high doses of venom may be associated with the lung alterations observed in humans. In fact, in different models of hemorrhagic shock, plasma, pulmonary and hepatic increases in IL-6 and MCP-1 were observed along with inflammation and lung injury, which may culminate in acute respiratory distress syndrome [97,98,99]. It is important to emphasize that in addition to cytokines, some other factors may be associated with the pulmonary hemorrhaging and death caused by a venom overdose. Further, the hemostatic alterations promoted by SVMP, SVSP and PLA 2 can also contribute to lung hemorrhage.
The treatment indicated for envenomation by N. annulifera is serum therapy. N. annulifera venom is part of the antigenic mixture used for the production of polyvalent serum by the South African Vaccine Producers (SAVP) (Pty) Ltd [28], although its immunogenicity and neutralizing potential have been poorly investigated. Here, we show that N. annulifera venom is highly immunogenic in murine model. Although with different intensities, this mouse antivenom was able to recognize venom components by Western Blot, mainly the ones with high molecular weight, which includes components as HYA, LAAO, CVF and SVMP. Moreover, this monovalent antivenom was able to protect the animals from death induced by venom, with high potency. In contrast, other studies have shown that horse antivenoms produced against venom mixtures, in which N. annulifera was included, were not able to neutralize the lethal effects of this venom [56,100,101]. This may be due to differences related to the animals used (mouse versus horse) or to a low level of neutralizing antibodies generated by other venoms present in the immunization pool.
In conclusion, here, we show that N. annulifera snake venom contains several components with toxic and pro-inflammatory properties. Some of these toxins promote coagulation disturbances, local and systemic inflammatory reactions, which may contribute to the pathologic events, observed in our murine model and possibly in dogs and humans envenomated by N. annulifera. Moreover, the venom promoted lung haemorrhage, an event that may also occur in cases of human envenomation, since death by respiratory arrest can be the result of the sum of neurotoxin activity and lung haemorrhage. High levels of IL-6 and MCP-1, as detected in the plasma of the envenomated animals, may be associated with pulmonary damage, since systemic inflammatory conditions can be deleterious and affect several organs, including the lungs. Thus, inflammation may be considered as target for the development of new therapeutic strategies in cases of N. annulifera human envenomation. Moreover, we showed that the venom is highly immunogenic and that the experimental serum was able to neutralize its lethal activity in the murine model. These data encourage further studies to characterize and produce monospecific therapeutic antivenom against N. annulifera. To evaluate the contribution of the different inflammatory mediators classes in hindpaw edema induced by N. annulifera' s venom, Balb/c mice groups were treated with [A] Cromolyn, a mast cell degranulation inhibitor (10 mg/kg, three days consecutively before the edema, i.p. route); [B] dexamethasone (Dexa), a cPLA 2 inhibitor (2 mg/kg, 2 hours before the edema, i.p. route); [C] indomethacin (Indo), a COX isoform inhibitor (10 mg/kg, 30 minutes before the edema, i.p. route); [D] MK886, a 5-lipoxygenase-activating protein inhibitor (5 mg/kg, 30 minutes before the edema, i.p. route) and [E] WEB2086, a PAFR antagonist (5 mg/kg, 1 hour before the edema, s.c. route). After treatment, edema was induced, and the paws were measured with a caliper rule at several time points. The increased paw volume was expressed as percentage. Statistical analysis was performed using two-way ANOVA, followed by the Bonferroni multiple comparison test ( ��� p� 0.05) ± SD. (TIF)