The venom of the scorpion Brotheas amazonicus (BamazV, following the systematic nomenclature proposed by Delgado-Prudencio and collaborators) (Delgado-Prudencio et al., 2025) was obtained by extraction of 17 specimens (7 males and 10 females) collected in Manaus, capital of Amazonas, Brazil (3°03′03.0″S 59°59′52.9″W). Briefly, the animals were subjected to 10 V electrical stimulation in the telson region, with subsequent lyophilization of the venom pool and storage at −80 C until the experiment. The experiments were authorized by the Ministry of the Environment under SISBIO (Authorization for Research in Federal Conservation Units) permit no. 56748–1 and SISGEN (National System for Management of Genetic Heritage and Associated Traditional Knowledge) registration no. A4A9FDD.

The breast tumor cell lines SKBr3, MCF-7, and MDA-MB-231, as well as the control breast epithelial cell line MCF10A, were obtained from the American Type Culture Collection (ATCC, Manassas, VA) and maintained in RPMI 1640 (Roswell Park Memorial Institute) supplemented with 10% FBS and 1% penicillin/streptomycin at 37 C, 5% CO₂. Cell death assay conditions were standardized using Paclitaxel on 4 cell lines. To standardize the cell death assay, 1 × 10⁵ cells of the 4 cell lines were seeded and treated 24 h later with Paclitaxel (0.01–1,000 µM). After 24 h of incubation, cell viability was measured using the MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay (Supplementary Figure S1). Different concentrations of venom (2–250 μg/mL), its >10 kDa, 3–10 kDa, and <3 kDa fractions (0.2–25 μg/mL), respectively, Bamaz>10, Bamaz3-10, and Bamaz<3. Obtained through venom ultrafiltration (Supplementary Figure S2), and selected peaks from the >10 kDa fraction eluted on the RP-C18 column (0.0004–40 μg/mL, including major peaks and those with enough protein for bioassays, regardless of their abundance) were used to establish the IC₅₀ curve (concentration required to achieve 50% of the maximum inhibitory effect) in cultures of healthy and tumor human cell lines. The samples were diluted in RPMI supplemented with 10% FBS and 1% penicillin/streptomycin and placed in culture for 24 h until analysis. RPMI was used as a negative control, and 30% DMSO was used as a positive control. Cells were seeded at a density of 1 × 10⁵ cells/well in a 96-well (Corning, 3,599) plate for stimulation. For all experiments, cells were left for 24 h for adhesion plus 24 h with stimuli before MTT or Flow Cytometry assays.

Bamaz>10 (7 mg, Supplementary Figure S2) was applied on a reversed-phase C18 column (10.0 mm × 250.0 mm, 5 μm, 300 Å, Jupiter^®, Phenomenex, United States) using a Fast protein liquid chromatography (FPLC) system. The components adsorbed onto the resin matrix were eluted using a step concentration gradient from 0% to 100% of solution B (80% acetonitrile (MeCN) in 0.1% trifluoroacetic acid (TFA)) at a constant flow rate of 5 mL/min over 240 min. The gradient profile included an initial plateau of 0% for 4 min, followed by linear increases from 0% to 25% over 36 min, 25%–60% over 140 min, 60%–100% for 40 min, a final plateau of 100% maintained for 12 min and re-equilibration at 0% for 8 min. Absorbance was automatically registered at 214 nm by the FPLC Äkta Purifier UPC-10 system (GE Healthcare, Sweden), and fractions of 2.5 mL/tube were collected at 25 °C. Peak 59 (50 µg) was rechromatographed on a 2.1 mm × 250.0 mm C18 column (5 μm, 300 Å, Jupiter^®, Phenomenex), eluting with a step concentration gradient from 0% to 100% of solution B, at a flow rate of 0.5 mL/min over 44 min. The gradient profile included an initial plateau of 0% for 8 min, followed by linear increases from 0% to 42% over 4 min, 42%–48% over 16 min, 48%–100% for 2 min, a final plateau of 100% maintained for 2 min and re-equilibration at 0% for 12 min. Absorbances (λ = 214 nm) were registered, and fractions of 0.5 mL/tube were collected at 25 °C.

Brotheas amazonicus crude venom (BamazV), its fractions (Bamaz>10, Bamaz3-10, Bamaz<3) (15 µg/well), and selected peaks (P9, P27, P28, P31, P32, P36, P57, P59, P75, P81) were dispersed in 20 µL of water and centrifuged at 13,000 xg, 4 °C, for 5 min. Protein concentration in the supernatant was estimated using a NanoDrop 2000 spectrophotometer (Thermo Scientific, United States) at 280 nm. Despite the limitations of UV absorbance for complex mixtures, this method was selected due to its low sample consumption and suitability for small-volume, low-yield venom fractions. A uniform extinction coefficient (ε₂₈₀ = 1.0) was applied for relative quantification, and all measurements were performed in technical triplicate to ensure consistency. Soluble samples (3–15 µg of total protein/well) were prepared in Laemmli sample buffer containing final concentrations of 2% SDS (w/V), 10% glycerol, 5% β-mercaptoethanol, 0.002% bromphenol blue and 0.125 M Tris HCl, pH 6.8, Samples were heat-denatured at 95 °C for 5 min to ensure complete denaturation and reduction of disulfide bonds. Proteins were then separated onto a 10% T, 6% C separating Tricine Sodium Dodecyl Sulfate Polyacrylamide Gel Electrophoresis (Tricine-SDS-PAGE), overlaid by a 5% stacking gel. The stacking gel was prepared by mixing 200 µL of AB-6 stock solution (49.5% T, 6% C), 667 µL of 3× Tricine gel buffer (3 M Tris-HCl, 0.3% SDS, pH 8.45), and ultrapure water to a final volume of 2 mL (Schägger, 2006). Differences in protein loading reflect variation in protein yield among fractions and peaks; although equal loading was prioritized for comparative analyses, some samples (e.g., P75) were limited by available material. Electrophoretic separation was initiated and maintained at 60 V for 20 min, followed by a subsequent adjustment to 10 W/gel until the end of the electrophoretic run (Jiang et al., 2016). For molecular mass estimation, two commercial protein standards were used: Low Range Molecular Weight Marker (Sigma M3546; 1.06–26.6 kDa) and Precision Plus Protein™ Dual Color Standards (Bio-Rad #1610374; 10–250 kDa), allowing accurate visualization across a broad mass range. Gels were stained with 0.025% Coomassie Blue G-250, and the image was captured using a Gel Doc™ EZ Imager and the software Image Lab version 5.2.1 (Bio-Rad, United States).

The viability of cells (MCF10A, SKBr3, MCF7, and MDA-MB-231) exposed to the fractions and/or peptides (crude venom, fractions <3, 3–10, and >10 kDa and peaks P9, P27, P28, P31, P32, P36, P57, P59, P75, P77, P81, AND P85) was determined by the MTT reduction method. After 24 h of treatment, the wells received 10 μL of MTT (5 mg/mL) and were incubated for 4 h at 37 °C. After incubation, plates were centrifuged at 400 g for 5 min, the supernatant was discarded by inversion. Formazan crystals were solubilized in 100 μL of DMSO (dimethyl sulfoxide) under orbital shaking for approximately 20 min. The absorbance of the wells at 570 nm was then read. The absorbance results of the control groups (treatment with culture medium only) corresponded to 100% cell viability, and the groups treated with peptides had their viability percentages calculated from the control. All the experiments were performed in triplicate and were independently repeated at least 3 times.

Cells were seeded at 1 × 10⁵ concentration in 96-well (Corning, 3,599) plates and left overnight for adhesion. The cells were treated with BamazScplp1 (50 μg/mL) for 24 h. The cells were then released from the plate and transferred to round bottom 96-well plate to start labeling with Annexin V (eBioscience, 88–8,103–74) and Fixable Viability Stain 780 (FVS) (eBioscience, 65–0865–14). For labeling, the cells were washed with Phosphate buffered saline (PBS) and labeled with FVS 1:1000 for 15 min at room temperature, after which the cells were washed with Annexin V Binding Buffer (Invitrogen) and labeled with Annexin V diluted in Annexin V Binding Buffer for 15 min at 4 °C protected from light. The samples were passed through the BD FACSCanto™ II equipment. Analyses were performed using FlowJo software version 10 (Becton Dickinson and Company, Franklin Lakes, NJ, United States). 10,000 events were acquired for all cells. Cells were gated by size (FSC) and granularity (SSC), singlets (FSC-A X FSC-H), followed by Annexin + FVS- (apoptosis) and Annexin-FVS+ (necrosis) (Supplementary Figure S6).

The initial 43 amino acid residues from the N-terminal region of the rechromatographed peak 59 were determined through Edman degradation (Edman and Begg, 1967) using an automated PPSQ-33A sequencer (Shimadzu Co., Japan). Sequence analysis was performed with BLAST (Basic Local Alignment Search Tool). Scorpine-like peptide sequences were retrieved from the Universal Protein Resource Knowledgebase (www.uniprot.org). Sequence alignment was created by the MultAlin Interface Page (Corpet, 1988), and the figure was generated by the ESPript server (Gouet et al., 1999).

Rechromatographed peak 59, hereafter designated BamazScplp1 (Bamaz-scorpine-like-1), was named following the systematic nomenclature proposed by Delgado-Prudencio et al. (2025) and the naming convention reported at UniProt entry P0C8W5 (Hg-scorpine-like-2). The sample was resuspended in TA50 (50% (V/V) MeCN in 0.1% (V/V) TFA), mixed at 1:1 (V/V) ratio with a saturated matrix solution of sinapinic acid (SA) in TA50, and subjected to matrix-assisted laser desorption/ionization (MALDI) time-of-flight (TOF) mass spectrometry analysis using an UltrafleXtreme system (Bruker Daltonics, United States). Mass spectra were acquired in positive reflector mode with either 500 or 2,500 laser shots per spectrum (single position), following external calibration with Bruker Protein Calibration Standard I (mass range: 4,000–20,000 Da; Bruker Daltonics). Mass spectrometry analyses (MS and MS/MS) were performed in two independent technical replicates derived from the same sample (P59), each spotted at randomized positions on the MALDI target plate. These replicates were used to confirm the analytical reproducibility of mass measurements and peptide fragmentation patterns. Data were processed using flexAnalysis software, version 3.4.76.0(Bruker Daltonics).

Coomassie-stained gel bands of P59 were excised from Tricine-SDS-PAGE gels and destained with 50% (V/V) methanol in 50 mM ammonium bicarbonate (pH 8). Samples were reduced with 10 mM dithiothreitol (DTT) at 56 °C for 40 min under agitation (600 rpm), followed by alkylation with 55 mM iodoacetamide at 25 °C in the dark for 1 h. The gel pieces were then dehydrated with acetonitrile and dried.

In-gel digestion was performed by incubating each sample with 40 ng of sequencing-grade modified trypsin (Promega V5111) in 4 µL of 50 mM ammonium bicarbonate, at 37 C for 15 h under continuous agitation (600 rpm). The digestion was stopped by adding 1% (V/V) TFA. Resulting peptides were desalted using ZipTip^® Pipette Tips with 0.6 µL C18 resin (MilliporeSigma, United States), and eluted in 50% acetonitrile containing 0.1% TFA.

For MALDI-TOF mass spectrometry analysis, desalted peptide samples were mixed at a 1:1 (V/V) ratio with a saturated solution of α-cyano-4-hydroxycinnamic acid (HCCA; 10 mg/mL in TA50) and spotted onto a polished steel target plate. Spectra were acquired in positive reflector mode with 1,500 laser shots per spectrum, following external calibration using the Bruker Peptide Calibration Standard (mono-isotopic, 1,000–3,500 Da; Bruker Daltonics). The method yielded a mass resolution exceeding 14,000 and a mass accuracy below 50 ppm. Data were acquired from multiple technical replicates at randomized target positions to ensure analytical reproducibility and robustness.

Fragmentation of the ten most intense precursor ions was performed by MALDI-TOF/TOF using post-source decay in LIFT mode without collision gas. Each fragmentation spectrum was generated from 2,500 laser shots.

Raw MS and MS/MS data were processed using MASCOT MS/MS Ions Search (https://www.matrixscience.com/cgi/search_form.pl?FORMVER=2& SEARCH=MIS) with a precursor mass tolerance of 1.2 Da and a fragment mass tolerance of 0.6 Da. Carbamidomethylation of cysteine was set as a fixed modification; and oxidation (Met/His/Trp), Gln- > pyro-Glu (N-term Q/N-term E), and deamidated (Asn/Gln) were considered variable modifications. Results were searched against the SwissProt database (taxonomy: ‘Metazoa’). Additionally, all spectra were manually investigated and manual de novo sequencing was validated using the MS-Product tool (https://prospector.ucsf.edu/prospector/cgi-bin/msform.cgi?form=msproduct).

All data represent mean ± standard deviation. Data were analyzed using one-way ANOVA with Bonferroni posttest for multiple groups or Student’s t-test for two-group comparisons. Data were considered statistically significant when p < 0.05. Information about replicates and independent experiments for each analysis are included in figure legends.

As a means of validating the chemosensitivity of the selected breast cancer cell lines, we first tested their response to paclitaxel at concentrations of 0.01–1,000 μM, a chemotherapeutic drug widely used in the clinical treatment of breast cancer. As shown in Supplementary Figure S1, all cell lines used responded in a dose-dependent manner to paclitaxel. All cell types responded to paclitaxel stimulation. MCF10A and SKBr3 responded at doses of 1 μM and above, while MCF7 and MDa-MB-231 responded at doses of 10 μM and above. In the next experiment, we used crude venom from the B. amazonicus scorpion at concentrations of 2, 10, 50 and 250 μg/mL (Figure 1). Most cells responded with death at 50 μg/mL and above, with the triple negative cell line (MDA-MB-231) being the most susceptible, responding at all venom doses. The Selectivity Index (SI) (IC₅₀ no cancer cell/IC₅₀ cancer cell) was 0.416, 0.5, and 2.5 for SKBr3, MCF7, and MDA-MB-231 respectively. This experiment demonstrated that crude venom from the B. amazonicus scorpion showed cytotoxic activity in tumor cells, comparable to paclitaxel, suggesting potential antitumor effects that warrant further investigation.

To refine our investigation and determine the molecular weight range of the molecule responsible for the cytotoxic activity observed in BamazV, we subjected the dispersed venom to an ultrafiltration process, which fractionated the sample into distinct molecular weight ranges, as shown in Supplementary Figure S2. Thus, we stimulated breast tumor cells with fractions >10 kDa, 3–10 kDa and <3 kDa derived from the crude venom. As observed in Figure 2, only the fraction >10 kDa maintained its cytotoxic activity, promoting the death of almost 50% of the cells with a dose of 25 μg/mL. These results demonstrate that the molecule with cytotoxic activity present in BamazV probably has a weight greater than 10 kDa. This finding indicates that the active component is likely a protein, which supports the application of protein-specific strategies for its purification and structural characterization. Initial isolation may involve chromatographic techniques or electroelution from SDS-PAGE if the activity can be associated with a specific band. Subsequent characterization could include mass spectrometry to determine molecular weight, sequence identification via LC-MS/MS or Edman degradation, and analysis of tertiary structure using circular dichroism, NMR spectroscopy, or other high-resolution methods.

Bamaz>10 was subjected to reversed-phase fast protein liquid chromatography (RP-FPLC) on a C18 column (250 mm × 10 mm). Its major peak, fraction P59 (Figure 3A), representing 14.3% of the total soluble protein in the >10 kDa fraction, was further purified by rechromatography on a different C18 column (250 × 2.1 mm). The resulting pure toxin, which constituted 79.7% of P59 (Figure 3B), was designated BamazScplp1 and accounted for 11.4% of the total protein in Bamaz>10. SDS-PAGE analysis revealed distinct protein banding patterns, indicating the presence of multiple proteins in the different eluted peaks (Figure 3C). The electrophoretic profile of Bamaz>10 closely resembled that of BamazV, suggesting a similar protein composition. The detection of proteins smaller than 10 kDa in the >10 kDa fraction indicates that the ultrafiltration process using a 10 kDa regenerated cellulose membrane (Amicon^® Ultra 15) did not achieve complete size-based separation. This limitation is consistent with the known variability around membrane cutoff thresholds and may be exacerbated by nonspecific adsorption or conformational flexibility of small proteins, which can facilitate their retention or passage through the membrane under centrifugal conditions. Nevertheless, the ultrafiltration step was intended as a preparative enrichment strategy, and the major cytotoxic component, BamazScplp1, was consistently recovered in the >10 kDa fraction, as confirmed by SDS-PAGE and mass spectrometry, supporting the reliability of the downstream analyses. Notably, no bands were detected in the Bamaz3-10 fraction, possibly due to a low sample amount or limited staining efficiency. P59 exhibited two protein bands within the 10–14.2 kDa range. Several peaks exhibited protein bands within the 6.5–14.2 kDa range, while P27 also contained bands above 14.2 kDa (Figure 3C). The lane of P75, loaded with 2.7 µg of protein, displayed at least four distinct protein bands above 37 kDa. In contrast, P57 showed no visible bands, despite loading approximately three times the amount used for P75, which was clearly detected in the gel. Similarly, no bands were observed for P9, even with a high protein load of 42 µg.

Following chromatographic separation of the >10 kDa fraction, we investigated the cytotoxic potential of individual peaks against breast-derived cell lines. As shown in Figure 4, four peaks exhibited significant cytotoxic effects. Peaks 31 and 81 selectively reduced the viability of the SKBr3 cell line at concentrations of 0.075 μg/mL and 0.015 μg/mL, respectively. In contrast, peak 75 demonstrated selective cytotoxicity only toward the non-tumorigenic MCF10A cell line at 0.04 μg/mL. The low activity levels exhibited by peaks P31, P75 and P81 may be due to high toxicity and selectivity, deserving further studies to elucidate their mechanism of action. Notably, peak 59 induced cytotoxicity across all tested tumor cell lines (SKBR3, MCF7, and MDA-MB-231), as well as the control cell line MCF10A (at 31 μg/mL or higher). Additional peaks–P9, P27, P28, and P32 (Supplementary Figure S3); P36, P57, P77, and P85 (Supplementary Figure S4) – showed no detectable cytotoxic activity under the tested conditions. Based on its broader spectrum of activity, peak 59 was selected for further characterization to elucidate its molecular properties and mechanisms of action related to cell death induction.

The mass spectrum of P59 showed a major peak intensity at an average ion m/z 9,199 Da ([M + H]⁺), along with an ion at m/z 4,604 Da ([M + 2H]²⁺; Figure 5A), designated BamazScplp1. Additionally, P59 included another peak at an average ion m/z 13,586 Da and an ion at m/z 6,793 Da ([M + 2H]²⁺; Figure 5B). The first 43 N-terminal amino acids from BamazScplp1 were determined by Edman degradation method as GLIKEQYFHKANDSLSYLIPKPVVNKLVGNA AHQMIHGIGGVQ. This N-terminal sequence corresponds to 50.9% of the molecule.

In addition, the internal peptide EQYFHK from BamazScplp1 (Supplementary Figure S5) was identified by de novo sequencing. Although additional peptide sequences could not be reliably deduced due to suboptimal fragmentation patterns, the complete mass fingerprint is provided in Table 1. The protein sequence data of BamazScplp1 reported in this paper will appear in the UniProt Knowledgebase under the accession number C0HME9. Sequence alignment revealed 46%–55% identity and 74%–81% similarity with scorpine-like sequences deposited in public databases, including: a putative scorpine-like peptide from Superstitionia donensis (A0A1V1WC79), Hg-scorpine-like-2 from Hoffmannihadrurus gertschi (P0C8W5), and scorpine-like-2 from Urodacus yaschenkoi (L0GCW2) (Figure 5C).

Considering the cytolytic activity of BamazV (Figure 1), we addressed the question whether BamazScplp1 would induce cell death by apoptosis or necrosis. All cell lines were treated with 50 μg/mL BamazScplp1 for 24 h (Figure 6). After treatment, the cells were assessed for viability and type of cell death. MCF10A cells showed significantly reduced viability after peptide exposure, with decreased apoptosis and a marked increase in necrosis compared to the untreated control. The same phenotype can be observed for the SKBR3 and MDA-MB-231 strains. The MCF7 strain showed reduced cell viability and increased necrotic cell death following treatment. However, no significant difference in apoptosis was observed compared to the untreated control. These findings indicate that necrosis is the major mechanism of cell death induced by BamazScplp1, consistent with reports in the literature describing other scorpine-like molecules with high cytolytic activity.