Antibacterial activity of six novel peptides from Tityus discrepans scorpion venom. A fluorescent probe study of microbial membrane Na + permeability changes

Toxicon 54 (2009) 802–817 Contents lists available at ScienceDirect Toxicon journal homepage: www.elsevier.com/locate/toxicon Antibacterial activity of six novel peptides from Tityus discrepans scorpion venom. A fluorescent probe study of microbial membrane Naþ permeability changes Patricia Dı́az a, Gina D’Suze a, *, Vı́ctor Salazar b, Carlos Sevcik a, John D. Shannon c, Nicholas E. Sherman c, Jay W. Fox c a b c Laboratory on Cellular Neuropharmacology, Biophysics and Biochemistry Center, Instituto Venezolano de Investigaciones Cientı́ficas (IVIC), Caracas, Venezuela Histology Service, Biophysics and Biochemistry Center, Instituto Venezolano de Investigaciones Cientı́ficas (IVIC), Caracas, Venezuela Biomolecular Research Facility, University of Virginia, Charlottesville, VA, USA a r t i c l e i n f o a b s t r a c t Article history: Received 18 June 2008 Received in revised form 4 June 2009 Accepted 8 June 2009 Available online 21 June 2009 Six novel peptides (named bactridines) were isolated from Tityus discrepans scorpion venom. From mass spectrometry molecular masses were 6916, 7362, 7226, 7011, 7101 and 7173 Da (bactridines 1–6). Bactridines 1 and 2 were sequenced by Edman degradation. The sequences and in silico analysis, indicated that they are positively charged polypeptides comprised of 61 and 64 amino acids (AA), respectively, bactridine 1 and bactridine 2 containing 4 disulfide bridges. Bactridine 1 was only toxic to cockroaches and crabs, and bactridine 2–6 were only toxic to mice. Bactridine 1 has a 78% sequence identity with ardiscretin. Ardisctretin is an insect specific sodium toxin which also produces a small depolarization and induces repetitive firing in squid axons resembling those of DDT [1,10(pchlorobenzyl) 2-trichloretane] in its ability to slow down action potential, to induce repetitive firing. Measured as the minimal inhibitory concentration, bactridines had high antibacterial activity against a wide range of Gram positive and Gram negative bacteria. Complete bacterial growth inhibition occurred at concentrations from 20 to 80 mM depending on the bacteria and peptide tested. Effects on membrane Naþ permeability induced by bactridines were observed on Yersinia enterocolitica loaded with 1 mM CoroNaÔ Red. CoroNaÔ Red fluorescence leakage from bacteria was observed after exposure to 0.3 mM of any bactridine tested, indicating that they modified Naþ membrane permeability. This effect was blocked by 10 mM amiloride and by 25 mM mibefradil drugs that affect Naþ and Ca2þ channels respectively. We found no evidence of changes of Kþ or Ca2þ concentrations neither inside nor outside the bacteria in experiments using the fluorescent dyes Fluo 4AM (10 mM) and PBFI (20 mM). Ó 2009 Elsevier Ltd. All rights reserved. Keywords: Scorpion venom Tityus discrepans Antibacterial activity Naþ channels 1. Introduction Living organisms are exposed daily to microbial infections. In order to defend themselves against the hostile environment, they have developed potent defensive * Corresponding author. IVIC CBB, Apartado 20632, Caracas 1020-A, Venezuela. Tel.: þ58 212 5041225; fax: þ58 212 5041093. E-mail address: gdsuze@ivic.ve (G. D’Suze). 0041-0101/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.toxicon.2009.06.014 mechanisms that are part of innate and adaptive immunity (Bulet et al., 2004). Antibacterial peptides are the effector molecules of innate immunity and the primary targets are bacterial membranes, (Boman, 2003). Antimicrobial peptides were initially discovered in invertebrates (Steiner et al., 1981) and later also have been described in vertebrates (Zasloff, 1987; Lehrer et al., 1991). These peptides show diverse sequences, structures and target specificity. Some of them are cationic, amphipathic peptides with P. Dı́az et al. / Toxicon 54 (2009) 802–817 molecular masses below 10 kDa and higher specificity for prokaryotic than for eukaryote cells (Matsuzaki et al., 1999). With the growing problem of pathogenic organisms resistant to conventional antibiotics, there is increased interest in the pharmacological application of antimicrobial peptides (Epand and Vogel, 1999). Antimicrobial peptides can be divided into three major groups: (1) amphipathic linear peptides free of cysteine and often with a-helical structure, (2) peptides with three disulfide bonds and (3) peptides with unusual abundance in certain amino acids, such proline, arginine, tryptophan or histidine (Boman, 2003). The first antimicrobial peptide isolated from scorpions were of the defensin type from Leiurus quinquestriatus hebraeus haemolymph (Cociancich et al., 1993). Later cytolitic and/or antibacterial peptides were isolated from scorpions belonging to the Buthidae, Scorpionidae, Ischnuridae, and Iuridae superfamilies’ hemo-lymph and venom (Ehret-Sabatier et al., 1996; Conde et al., 2000; Torres-Larios et al., 2000; Corzo et al., 2001; Dai et al., 2002; Moerman et al., 2002; Rodriguez de la Vega et al., 2004; Uawonggul et al., 2007). The discovery of these peptides in venoms from Eurasian scorpions, Africa and the Americas, confirmed their widespread occurrence and significant biological function. The main site of antimicrobial activity is the plasma membrane of bacteria and parasitic protozoans. Some antimicrobial peptides are ionically attracted to negatively charged groups of the cell surface, where they adopt an a-helical conformation and accumulate on the membrane. This may result in the formation of transient pores, membrane perturbation and cell lysis (Giangaspero et al., 2001). Other antimicrobial peptides have been shown to induce bacterial membrane permeability perturbations indicating interaction with membrane ion channels (Hetru et al., 2000). Their mechanism of action on bacteria membrane ion channels is not well understood. In the present paper, we report the isolation of six antibacterial peptides from T. discrepans scorpion venom. Their ability to induce bacterial membrane permeability changes was studied with epifluorescent methods under several experimental conditions. The biological activity and pharmacological interactions of these peptides with tetrodotoxin, amiloride and mibefradil on bacterial Naþ-channels, suggest that they act on channels resembling the NaChBact channels of Bacillus hallodurans (Ren et al., 2001). Some of the results presented here were previously communicated in preliminary form (Dı́azBello et al., 2007). 803 2.2. Isolation procedures Soluble venom (100 mg each time), was applied to a Sephadex G-50 column (200  1 cm) as described previously (D’Suze et al., 1995). Fraction II was collected, freezedried and submitted to an analytical C18 reversed phase column (250  10 mm) (Vydac, Hesperia, CA, USA). Components were separated using a linear gradient from solution A [0.12% trifluoroacetic acid (TFA) in water] to 45% solution B (0.10% TFA in acetonitrile), in 45 min, at a flow rate of 1 mL/min. The elution was monitored by absorbance at 230 nm. Active peaks were re-chromatographed through the same column, using a linear gradient from 25 to 35% solution B in 60 min at a flow rate of 1 mL/min. Peaks were collected manually, dried using a Savant Speed-Vac dryer and used for mass spectrometry analysis, automatic Edman degradation, minimal inhibitory concentration, and for fluorescence experiments. 2.3. Bacterial strains The anti-microbial activity of the isolated peptides were tested against a set of three Gram positive species: Bacillus subtillis (ICTA-07), Microccocus luteus (ATCC4698), Enterococcus faecalis (WHO14), and three Gram negative species: Pseudomonas aeruginosa (ATCC27853), Y. enterocolitica (ATCC9610) and Acinetobacter calcoaceticus (ATCC2305-5), purchased from Centro Venezolano de Colecciones de Microorganismos (Caracas, Venezuela). Bacteria were grown aerobically at 37  C in LB medium (Sigma-Chemical, USA) except Y. enterocolitica which was grown in Brain Heart Infusion (Oxoid, Hampshire, England) and A. calcoaceticus in Nutrient Broth (Himedia, Mumbai, India). 2.4. Antibacterial activity Antibacterial activity was determined using the microdilution assay. Bacteria were grown at 37  C in LB medium, Brain Heart Infusion or Nutrient Broth for 6–7 h (until the exponential phase), to approximately 0.8 UA620 which corresponds to 3.2  108 colony forming units per mL (CFU/ mL). Then the CFU were diluted in the same medium to approximately 0.002 UA620 corresponding to between 106 and 107 CFU/mL. Bacteria were incubated in 96-well microplates (Nunc, Denmark) in presence of different peptides concentrations, at 37  C; after 16 h, A620 was measured. Minimal inhibitory concentration (MIC) was defined as the lowest concentration causing 100% grown inhibition. 2. Materials and methods 2.5. Mass spectrometry analysis and Edman sequencing 2.1. Venom source 2.5.1. MALDI-TOF mass spectrometry Samples were initially subjected to MALDI-TOF mass spectrometry. Samples were dissolved in 0.1% trifluoroacetic acid. For mass determination, 0.5 mL of sample was spotted on a plate followed by addition of 15 mg/mL sinapinic acid in 50% acetonitrile, 0.1% trifluoroacetic acid. The sample was analyzed in a Bruker Microflex time of flight mass spectrometer or Applied Bioystems Voyager DE time of flight mass spectrometer using the manufacturers’ T. discrepans scorpions were collected in the metropolitan area surrounding Caracas, Venezuela. Scorpions kept alive in the laboratory were anesthetized once a month with CO2 and milked for venom by means of electrical stimulation. Venom was dissolved in double distilled water and centrifuged at 15,000 g for 15 min. The supernatant was freeze-dried and stored at 80  C until used. 804 P. Dı́az et al. / Toxicon 54 (2009) 802–817 protocols. A mixture of small proteins was used to calibrate the instrument which was operated in linear mode to determine average mass to charge ratio. 2.5.2. LTQ-FT mass spectrometry Some of the peptide masses were checked also with a linear quadrupole ion trap Fourier transform cyclotron resonance mass spectrometer (Thermo Scientific LTQ-FT). The samples in 50% acetonitrile 0.5% acetic acid (about 25 mL) were infused into the LTQ-FT MS at a rate of 0.3 mL/ min through a tip manually fabricated from 75 mM i.d.  360 mM o.d. fused silica (Polymicro Technologies). The electrospray voltage on the LTQ was set to 3.5 kV, and the tube lens voltage was 100 V. The temperature on the capillary was 200 degrees. Once a consistent signal was obtained, 50 FT scans were collected and averaged over a mass of 300–1600 m/z, at a resolution of 100,000. The charge state of the species detected was determined by the spacing of the isotope peaks. The monoisotopic peak was identified using the averaging method (Senko et al., 1995). Through a simple calculation using the m/z of the monoisotopic peak and the charge state of the peak, the monoisotopic mass of the species was determined. 2.5.3. Edman sequencing Edman sequencing was performed on an Applied Biosystems Procise 494 sequencer. A Biobrene (Polybrene) coated glass fiber filter was cleaned with the manufacturer’s cycles, then the samples were loaded on the filter and dried before analysis using the manufacturer’s pulsed liquid cycles. Amino acids were identified by chromatography with the standard Procise chromatography system and protocols. 2.5.4. In silico tertiary structure prediction methods The sequences of bactridines 1 and 2, were used to predict their tertiary structure by molecular homology simulation. For this purpose their amino acid (AA) sequences were submitted to SwissModel BlastP 2.9.9 template identification server (Altschul et al., 1997) to search for adequate templates. Through this procedure 1npi motif A (T. serrulatus Ts1 neurotoxin, Pinheiro et al., 2003) was selected to model the peptides. Each sequence was then submitted to the SwissModel server in automatic mode with the chosen template to obtain a tertiary structure model (Guex and Peitsch, 1997; Schwede et al., 2003; Arnold et al., 2006). The sequence identity between the chosen templates and the bactridine sequences were within the, so called, ‘‘safe zone’’ for homology modeling, so that similar sequences adopt practically identical structures, and distantly related sequences still fold into similar structures (Chothia and Lesk, 1986; Sander and Schneider, 1991; Rost, 1999). Disulfide bonds for the AA sequences were also predicted with the DISULFIND server (Ceroni et al., 2006). Missing AA in the pdb file returned by the SwissModel server or missing disulfide bonds, were completed with the YASARA molecular graphics, modeling and simulation program (www.yasara.org) under Ubuntu Linux v 8.10. Side chains of the modeled molecules were reoptimized with the molecular modeling program YASARA, such that the YAMBER3 force field energy was minimal and binding energies could be calculated (Krieger et al., 2004). Temperature was set at YASARAS default 298 K and pH 7. 2.5.5. Microscopic observation and integrated optical density measurements in bacteria Observation of fluorescent bacteria and imaging was done as Iwamoto and Allen (2004) with some modifications. The observations were done using an Eclipse E600 (Nikon, Tokyo, Japan) fluorescence microscope, equipped with epifluorescence illumination and a G-2A filter for CoroNaÔ red. Pictures of fluorescent images were taken with an 8 Mpixel Coolpix 8700 digital camera (Nikon, Tokio, Japan). With the bacteria attached to a coverslip placed on top of a Neubauer chamber and using a Hamilton microsyringe with a catheter on the tip connected to the chamber, 5 mL of 0.3 mM bactridine solution were added to the chamber. Fluorescence was measured as integrated optical density (IOD) inside the bacteria and its neighbouring medium from the digital photo image using ImageJ 1.38 software (http://rsb.info.nih.gov/Ij) (Ahn and Basbaum, 2006). Micrographies were taken immediately before and 0.25, 1, 5 and 10 min after adding bactridine to the chamber. Integrated optical density was measured in 10 randomly selected bacteria from each micrography to quantify fluorescence. Micrographies were also taken in the chamber with bactridine, but in the absence of bacteria, with the coverslip with poly-L-lysine using the same condition of illumination and same camera settings as done in presence of bacteria; then the background level of these micrographies was subtracted from the pictures containing bacteria. Bacterial width, length and surface area were measured in 50 randomly selected bacteria. The central cytoplasmic region from different cells was measured (before and 0.25, 1, 5 and 10 min after adding bactridine), using a selection ‘‘circle’’ (‘‘Elliptical’’ selection tool option in ImageJ) of 625 square pixels or 0.989 mM2, which corresponds to z31.1% of the bacterium surface. The selection ‘‘circle’’ occupied almost completely the central zone of bacteria (‘‘circle’’ diameter 1.122 mM, bacterium width 1.212  0.177 mm). The selection ‘‘circle’’ was then placed tangent to the surface of the widest bacterium part to measure external medium IOD. The IOD of these areas was automatically calculated by ImageJ, and represents the average of red, green, and blue color values (RGB) of analyzed area, expressed in optical units per surface area. The same procedure was applied to obtain the background optical density at each time in the bacteria-free pictures under the conditions explained above. A single area was enough for background subtraction, since the background was constant in each photomicrograph. The linearity of the measurement corresponding to the actual fluorescence under the experimental conditions was verified measuring IOD excited by different intensities of UV through the ND filters (Iwamoto and Allen, 2004). 2.5.6. Ion efflux measurements Y. enterocolitica at the mid-logarithmic growth phase was centrifuged at 120 g for 15 min to form a pellet. Cells were resuspended and washed twice with either double 805 P. Dı́az et al. / Toxicon 54 (2009) 802–817 distilled sterile water (DDSW) or sterile 0.15 M choline chloride in water, and incubated with CoroNaÔ Red (1 mM, 15 min), Fluo 4AM (10 mM, 15 min) or PBFI (20 mM, 40 min) (All dyes from Molecular Probes, Oregon, USA) at 37  C. CoroNaÔ Red is a cationic dye that can be loaded into cells as a fluorescent indicator of Naþ (Baron et al., 2005; Bernardinelli et al., 2006). Fluo 4AM is a fluorescent dye specific for Ca2þ (Gee et al., 2000) and PBFI is specific for Kþ (Ješek et al., 1990). Loaded bacteria were washed twice with DDSW or 0.15 M choline chloride, bound to glass cover slips with poly-L-lysine and placed into a homemade holder containing 5 mL DDSW or 0.15 M choline chloride. A holder carrying loaded bacteria was placed on top of Neubauer counting chamber on the microscope stage. Bacteria were exposed to bactridine 1–6 (0.3 mM) solutions. CoroNaÔ red fluorescence was measured as indicated above. To evaluate the effect of bactridines on Naþ channels, CoroNaÔ Red loaded bacteria were incubated with amiloride (10 mM), tetrodotoxin (30 mM), or mibefradil (25 mM) for 10 min and then exposed to bactridines (0.3 mM) mixed with the concentration of inhibitor under test. Loaded bacteria exposed to DDSW were used as control. 2.5.7. Toxicity assays Bactridines 1–6 were assayed in duplicate for toxicity in mice (IVIC strain, male, z15 g), by i.p. injection with 3 mg bactridine/g mice. Cockroaches (Periplaneta americana, z1 g) were injected through the leg first joint with 20 mg bactridine/roach. Control animals received saline (NaCl 0.9%). Animals were observed during 48 h to determine toxicity. Bactridines were also assayed in crabs (Uca rapax) of 5.75 (5.1, 6.6) g (n ¼ 37 crabs) body weight, two replicates per bactridine unless otherwise stated were used; bactridines were injected to crabs through the first joint of the legs. 2.5.8. Haemolytic activity assay Haemolytic activity of bactridines was evaluated by two different methods. It was evaluated as according to Kondejewski et al. (1996) with slight modifications. Human blood with heparin was centrifuged (1500 g) for 15 min at 4  C, were washed 3 times with isotonic NaCl saline (0.9%). The pellet was diluted to hundred times its volume and incubated with bactridindes 1–6 (0.15 mg/ml or 6 mM) for 10 min at 37  C. Total cell lysis was evaluated by incubation with 10% SDS in NaCl 0.9%, cells not exposed to bactridines in NaCl 0.9% were used as control. Lysis was determined visually using a low power stereomicroscope. Haemolysis was also evaluated as indicated by Chang et al. (1978) and Wishart et al. (1995) with slight modifications. Human blood was centrifuged (1500 g) for 15 min at 4  C, and the packed cells were washed 3 times with isotonic NaCl saline (0.9%). Two concentrations of each peptide (90 and 180 mM), were tested after incubating at 37  C for 30 min in 0.15 ml of 10% v/v erythrocyte in physiological solution. The cells were then centrifuged for 5 min at 1500 g, the supernatants transferred to a flat-bottom 96-well polystyrene microtiter plate, and the absorbance read at 540 nm in a microplate reader (Multiskan Labsystem). Controls for zero haemolysis (blank) and 100% haemolysis were supernatants of red blood cells suspended in physiological solution and 1% Triton X100 in isotonic NaCl saline (0.9%), respectively. Percent haemolysis was determined twice for each bactridine and its value calculated using the formula of Travis et al. (2000) as follows: %Haemolysis ¼ 100$ Asample  Ablank : Atriton  Ablank (1) Where: Asample is the plasma sample’s absorbance, Ablank is the b absorbance and Atriton is the Triton X100’s solution absorbance. 3. Results 3.1. Antibacterial toxin isolation T. discrepans venom produced five fractions in Sephadex G50 molecular exclusion gel chromatography but only one fraction (labeled with an asterisk, inset in Fig. 1A) inhibited bacterial growth. This fraction contains 73% of the venom eluted through the column. Fig. 1A shows the chromatographic profile of the active fraction on a reverse phase C18 analytical column. Twenty eight different peaks were eluted from minutes 20–49. Six components showed antibacterial activity and were named bactridines 1–6 (labeled with asterisks in Fig. 1A). The retention times (in min) of these fractions (listed in the naming sequence of the peptides, i.e.: bactridines 1 through 6) were: 36.45, 33.36, 33.86, 35.44, 35.54 and 32.75. After 3 re-chromatographies on C18 the purified peaks shown represent a yield of about 6.1, 1.7, 1.8, 1.8, 0.5, 0.5%, respectively, of the venom’s total protein content. This yields are based on the assumption that reading the absorbance at 280 nm, the specific absorption coefficient, called here a0 , equals 1 units of absorbance$ml solution$(cm path length)1$(mg 1 peptide) (expressed without units, from now on) for all the peptides. Fig. 1B, shows purified bactridines 1–6 as symmetric peaks obtained in the final re-chromatography. 3.2. Structural features of bactridines 3.2.1. Mass spectrometry and sequencing results The masses of bactridines 1–6 were determined using MALDI-TOF mass spectrometry as 6921, 7363, 7226, 7011, 7101 and 7173 Da, respectively. The MALDI-TOF values under this experimental conditions have an error of approximately 0.1% (7 Da). Bactridines 1 and 2, were also analyzed by high resolution, high mass accuracy via LTQ-FT mass spectrometry and were 6916 and 7362 Da, respectively. Amino acid sequences (Table 1) of bactridines 1 and 2 were determined by Edman degradation; in both cases a single amino acid at each degradation cycle was obtained suggesting a high level of purity. Cysteines gave no peaks in the direct sequencing procedure. Edman sequence analysis of unalkylated bactridine 1 allowed for residues 1–49 to be clearly identified except for positions 11, 15, 23, 27, 37, 42 and 44 which based on the presence of dehydroalanine were considered unprotected cysteines. Comparison with homologous sequences in the databases also suggests that this residues are cysteines. Residues 50–60 gave smaller peaks at the sequencer, but 806 P. Dı́az et al. / Toxicon 54 (2009) 802–817 Fig. 1. Chromatographic steps to isolate bacridines. Panel A: the inset is the elution pattern of T. discrepans scorpion venom on a column of Sephadex G-50 column (200  1 cm) as described previously. The fraction between arrows labeled with an asterisk was collected, freeze-dried and submitted to an analytical C18 reversed phase column (250  10 mm) (Vydac, Hesperia, CA, USA). The main part of panel A shows the components separated using a linear gradient from solution A [0.12% trifluoroacetic acid (TFA) in water] to 45% solution B (0.10% TFA in acetonitrile), for 45 min, at a flow rate of 1 mL/min. The elution was monitored by absorbance at 230 nm. Active peaks are labeled with an asterisk and a number identifying each bactridine (*1 is bactridine 1, *2 is bactridine 2, and so on). Panel B: the active peaks identified with asterisks and numerals in panel A, were re-chromatographed through the same column, using a linear gradient from 25 to 35% solution B for 60 min at a flow rate of 1 mL/min. Other details in the text of the communication. nevertheless sufficient for reasonable identification to be made. Therefore the sequence is the one shown on Table 1, where all cysteines are in disulfide bonds, giving the calculated mass of 6919 Da compared to the experimentally determined mass. In bactridine 2, prior to Edman sequence determination cysteines were alkylated with 4-vinylpyridine thus allowing for direct identification of the cysteines. From the Edman sequencing sixty three amino acids were identified and the last amino acid was surmised from mass spectrometry, giving the sequence presented in Table 1, where again all cysteines are proposed to be in disulfide bonds. Calculated molecular mass of the peptide is 7362 Da compared to the LTQ-FT-mass spectrometric determined mass of 7362 Da. In both cases the calculated masses of bactridine 1 and 2 from the Edman sequence analysis is well within the error for the masses determined by MALDITOF and LTQ-FT-mass spectrometry. 3.2.2. Bactridine 1 homology modeling We also investigated bactridine 1 using homology modeling. (see Methods for details not described here). The P. Dı́az et al. / Toxicon 54 (2009) 802–817 807 Table 1 Amino acid sequence of bactridines 1 and 2. homology modeling of bactridine 1 using 1npi motif A (T. serrulatus Ts1 neurotoxin, 59% identity between 1npi and bactridine 1 AA sequences) as template with the SwissModel server produced a pdb file including the first 60 AA of the peptidic sequence and no H atoms. All H atoms, as well as Cys61 were added by hand using YASARA; after adding Cys61 the molecule was optimized for minimum free energy. The SwissModel file for bactridine 1 had disulfide bonds between C15–C37, C23–C42 and C27–C44; after adding C61, its S atom was 2.007 Å from the S of C11 (a disulfide bond is z2.045 Å long) and a disulfide bond was established between them using YASARA, after which free energy was again minimized. A disulfide bonding scheme such as C11–C61, C15–C37, C23–C42 and C27–C44 was also predicted as the most likely one for this peptide using DISULFIND. Finally, the modeled structure was left to oscillate at 298 K until the YAMBER3 self parametrising force field energy of YASARA was minimal, the resulting 3D arrangement is presented as Bact 1 in Fig. 2. Calculated with YASARA bactridine 1 has a net charge of 0.456 at pH 7. Homology modeling was also used to infer on the relevance of disulfide bonds on bactridine 1 structure, the results are presented in the lower part of Fig. 2, at left is a cartoon depicting the tertiary structure of this peptide and its disulfide bonds (Bact 1 oxid, in the figure). The structure at the bottom left of the figure was produced using the Simulation feature of YASARA with the YAMBER3 force field, after, first, deleting the disulfide bonds, second, adding H to the freed sulfurs in the cysteines, and finally letting the simulation run for a while until the tertiary structure seemed to reach a stable form. When comparing Bact 1 oxid with Bact 1 red in Fig. 2, it becomes apparent that the disulfide bridges are critical to stabilize the tertiary structure of the peptide and in all likelihood also for its biological activity. 3.2.3. Bactridine 2 homology modeling The tertiary structure of bactridine 2 also was modeled using 1npi motif A as template (T. serrulatus Ts1 neurotoxin, 65.9% identity between 1npi and bactridine 2 AA sequences). In this case the SwissModel pdb file excluded also all H atoms and the Cys62, Gly63 and Arg64 segment of the C-terminus. Only two disulfide bridges were included in the SwisssModel file: C23–C43 and C27–C45, but the S atoms of C15 and C37 were left only 4.823 Å apart and a disulfide bond was established between them using YASARA as explained before. Finally, the three AA of the C-terminus were added using YASARA followed in each case by a free energy minimization cycle, the resulting 3D arrangement is presented as Bact 2 in Fig. 2. Just after adding Cys61, sit S atom was at 2.869 Å from the S atom of Cys11, and the two were linked using YASARA prior to adding the last two residues of bactridine 2 C-terminus. The disulfide bridge scheme (C11–C62, C15–C37, C23–C43 and C27–C45) shown for bactridine 2 in Table 1 was also the most likely scheme for disulfide bonds predicted by DISULFIND. Calculated with YASARA bactridine 2 has a net charge of 1.368 at pH 7. 808 P. Dı́az et al. / Toxicon 54 (2009) 802–817 Fig. 2. Homology modeling of bactridines 1 and 2. The tertiary structures of bactridines 1 (Bact 1) and 2 (Bact 2) calculated by homology modeling is presented in the top of the figure, and a cartoon representing the tertiary structure of bactridine 1 with disulfide bridges shown as ball and sticks (Bact 1 oxid) and Cys S reduced (Bact 1 red). Bact 1 and Bact 1 oxid in the atoms or residues that are presented as balls and sticks; to calculate Bact 2 red, the bonds between the S atoms shown in Bact 2 oxid were deleted in YASARA and the molecule was let to oscilate again using YASARA’s YAMBER3 force field until the resulting structure stabilised. In the top panes positive amino acids apear in blue, negative amino acids in red. Arg64 in the C-terminus of Bact 2 is grey, to indicate that the ionised carboxyl and guanidinium groups cancel each other to produce a residue with net charge 0; likewise Cys61 in Bact 1 is red to indicate that it has a negative charge due to its ionised carboxyl group. The S atoms in the lower panels of the figure have the color which correspond to the secondary structure where the Cys is located. In all 4 panels the secondary structure es colored as: blue, a-helix; green, b-turns; red, b-sheets; cyan, unstructured segments. See the text of the communication for modeling details. 3.2.4. Other in silico analyses The theoretical isoelectric points were determined as 8.17 and 8.99 for bactridines 1 and 2, respectively; likewise the values of a0 (for absorbance at 280 nm) were calculated as 3.314 assuming that all Cys appear as half cystines in bactridine 1, and were calculated as 3.720 assuming that all Cys appear as half cystines in bactridine 2 (http://expasy. org/to-ols/protparam.html; Gasteiger et al., 2005). The preceding values of a0 , calculated assuming that all Cys are forming cystines were used to calculate masses of bactridines 1 and 2 through this paper: the average of both values, a0 ¼ 3.447, were used to calculate masses of bactridines 3–6, whose peptidic sequences are currently unknown. 3.2.5. Antibacterial activity Bactridines’ antibacterial activities were assayed against Gram positive and Gram negative bacteria and expressed as MIC. As indicated above, from the pepitidic sequence of bactridine 1 and 2 their a0 were estimated as 3.242 and 3.720, respectively. The full sequences of the other bactridines are not yet known, but Table 2 shows that if a0 for the other bactridines is approximated by the average a0 of bactridines 1 and 2 (average ¼ 3.481), the peptides inhibited the growth of all bacteria studied, at concentrations between 20 and 72 mM. Table 2 Bactridines’ bacterial growth minimal inhibitory concentrations. Bacterium Bact 1 Bact 2 Bact 3 Bact 4 Bact 5 Bact 6 B. subtillis M. luteus E. faecalis P. aeruginosa Y. enterocolitica A. calcoaceticus 22 43 34 77 49 43 30 27 65 54 46 27 (35) (41) (41) (46) (35) (20) (32) (51) (46) (41) (26) (32) (20) (46) (51) (35) (26) (20) (23) (41) (61) (41) (41) (20) The values without parentheses were calculated setting the absorbance per unit path length (a0 ) as 3.242 for bactridine 1 and as 3.720 for bactridine 2. Values between parentheses were calculated assuming that a0 was 3.481 (Average of the a0 values of bactridines 1 and 2). All concentrations are in mM, see Results section for other details. P. Dı́az et al. / Toxicon 54 (2009) 802–817 3.2.6. Naþ efflux measurement Bactridines’ ability to induce bacterial cytoplasmic Naþ changes was studied in Y. enterocolitica loaded with 1 mM CoroNaÔ Red, under several experimental conditions. Fig. 3 shows fluorescent Y. enterocolitica pictures before and after adding bactridines 1 and 2 (1 mM mM). Bactridines induced an outflow of Na–CoroNa red complex from bacteria with a consequential reduction of intracellular fluorescence and a concomitant increase in extracellular fluorescence. This 809 outflow started 0.25 min after bactridine addition. Ten minutes later, the bacteria had lost almost all their intracellular fluorescence signaling changes on membrane Naþ permeability. As shown, amiloride and mibefradil prevented sodium outflow from the cells, and tetrodotoxin (TTX) had no effect. All 6 bactridines induced Naþ outflow from bacteria, and in no case this outflow was blocked by TTX, and also in all cases the outflow was blocked by amiloride. Fig. 4 Fig. 3. Micrographies of Y. enterocolitica loaded with the Na fluorescent indicator CoroNaÔ red (1 mM). The columns represent pictures of a bacterium at different times (indicated in minutes on top of each column). The different experimental conditions are indicated at left, the grey color is fluorescence of Na-CoroNa red complex. Water row: A bacterium not exposed to any drug retains the Na-CoroNa red complex in its cytoplasm for at least 10 min. Bact. 1 row: In a bacterium exposed to 0.3 mM bactridine 1 the grey color, initially confined to the cytoplasm of the bacterium, is observed to leak out as the time of exposure to the bactridines progresses. Bact. 2 row: Like the bact. 1 row, but using bactridine 2 instead. Bact. 1 þ Amiloride row: Like the bact. 1 row, but in a bacterium preincubated with 10 mM amiloride; as seen the grey color remains confined in the cytoplasm for at least 10 min. Bact. 1 þ Mibefradil row: Like the bact. 1 row, but in a bacterium preincubated with 25 mM mibefradil; as seen the gray color remains confined in the cytoplasm for at least 10 min. Bact. 1 þ TTX row: Like the bact. 1 row, but in a bacterium preincubated with 30 mM tetrodotoxin; as seen this potent blocker of electrically excitable Na-channels does not preclude the outflow of Na-CoroNa red complex. Other details in the text of the communication. 810 P. Dı́az et al. / Toxicon 54 (2009) 802–817 shows IOD (see methods) inside and in the immediate vicinity of CoroNaÔ red-loaded Y. enterocolitica exposed to bactridines 1, 2 and 4 (-), as well as the effects of amiloride (:) and mibefradil (C) blocking bactridine’s effect. As shown, all these bactridines produced an outflow of sodium which was completely inhibited by both amiloride and mibefradil. The figure also shows that under the effect of the bactridines 1, 2 and 4 IOD increased in the vicinity of the bacteria, and that this increase was blocked completely by both amiloride and mibefradil. Fig. 5 is similar to Fig. 4 except in that the effects of bactridines 3, 5 and 6 on CoroNaÔ red-loaded Y. enterocolitica and its vicinity are shown. These 3 bactridines also produced a decrease of intracellular IOD inhibitable by amiloride, but which was only partially inhibited by mibefradil in the case of bactridines 3 and 6, and not inhibited at all by this blocker in the case of bactridine 5. 3.2.7. Naþ efflux measurement in 0.15 M choline chloride Y. enterocolitica is an Enterobacteriaceae pathogen to humans and other animals, it survives in very different conditions ranging from tap water to body fluids. The experiments described in the previous section were carried out in distilled water, close to the lower extreme of the bacteria living conditions. To check bactridine action under osmolarity and ionic strength closer to body fluids, the action of bactridines 1 (;) and 2 (:) on Y enterocolitica sodium outflow were studied in a 0.15 M choline chloride water solution. As shown in the top panels in Fig. 6, there was a very large IOD increase in the vicinity of the bacteria exposed to the two bactridines, and this increase was largest when bactridine 1 was used. The IOD inside the bacteria, however, did not change significantly under the effect of neither of the two bactridines; yet, since we did not use confocal or deconvolution microscopy, the IOD increase in the vicinity on top and below the bacterial soma, probably compensated and overshadowed the decrease in IOD inside the bacteria. 3.2.8. Lack of effect of bactridines 1 and 2 on bacterial Kþ or Ca2þ outflux To determine whether the sodium outflow induced by the bactridines in CoroNaÔ Red-loaded Y. enterocolitica was a product of an unspecific collapse of the bacterial membrane, we studied the effect of the peptides in bacteria loaded with either Fluo 4AM (specific for Ca2þ) or PBFI (specific for Kþ) fluorescent dyes in 0.15 M choline chloride solution. As seen in Fig. 6, neither the IOD associated with Fluo 4AM (middle panels) nor with PBFI (bottom panels) changed, inside or outside the bacteria. 3.2.9. Toxicity assay Only bactridine 1 (20 mg/animal) induced excitability and death of cockroaches within 48 h after injection but it was not toxic to mice. Bactridines 2, 3, 4 and 6 (3 mg/g) caused excitability, sialorrhoea (excessive salivation), lacrimation, polyuria, diarrhoea, fasciculation and death in mice. Bactridine 5 (3 mg/g) had a transient excitatory effect in mice, but none of the injected animals died with the concentrations tested. We did not try to increase the bactridine 5 concentration due to ethical and cost considerations. Bactridine 1 at a dose as low as 0.9 mg/g induced sialorrhoea in crabs, when the dose was raised to 1.8 mg/g the sialorrhoea was more intense and 3 out of 7 crab died. At a dose of 1.8 mg/g, after 48 h of injection, bactridines 2, 3, 4, and 5 had no effect whatsoever in the crabs. Batridine 6 at doses as 0.9 and 1.8 mg/g produced sialorrhoea and hyperactivity in the crabs lasting z15 min, afterwards they became gradually depressed and loosed any defense bahaviour to the point they could be easily grabbed and did not try to pinch. 3.2.10. Haemolysis assay When bactridines 1, 3 and 5 were studied for haemolysis with the microscopy method of Kondejewski et al. (1996) no haemolysis of human erythrocytes after 10 min incubation occurred, but bactridines 2, 4 and 6 were haemolytic under these conditions. The results obtained measuring haemoglobin absorbance were also consistent with the microscopical method even at much higher concentration of bactridines. The results are summarized in Table 3, the percent haemolysis estimated for bactridines 1, 3 and 5 were 1% with any of the concentrations tested. From the percent haemolyses observed with these bactridines (close to zero and to the method’s resolution) no concentration dependence was clear. Bactridines 2, 4 and 6, in turn, had a large haemolytic effect which was concentration dependent. 4. Discussion Most of the antimicrobial peptides isolated from arthropod venoms are a-helical linear peptides. However two scorpion antimicrobial cysteine rich peptides, scorpine (Conde et al., 2000) and BmTXKS2 (Zhu et al., 2000), are structurally different. A new and unique antibacterial peptide which is also rich in cysteines and features of defensins and Kþ-channel blocking toxins is aurelin, recently isolated from the jelly fish Aurelia aurita (Ovchinnikova et al., 2006). In this work we have studied 28 peaks from T. discrepans venom Sephadex G50 fraction II, from these, only 6 were antibacterial, and were named bactridines. The structures of bactridines 4–6 are currently unknown but may possibly be similar to bactridines 1 and 2 based on chromatographic behavior, molecular mass, and antibacterial activity. Bactridines 1 and 2 were completely sequenced and found to be cysteine rich peptides like scorpine. However scorpine is able to block potassium membrane channels while bactridines act on sodium membrane channels. Bactridine 1 is 78% identical with ardiscretin a Naþ-channel insect b sodium toxin from T. discrepans venom (D’Suze et al., 2004) and 63% with TsNTxP a non-toxic protein from T. serrulatus venom (Guatimosim et al., 1999). Bactridine 1 was toxic to cockroaches and crabs, and non-toxic to mice. Table 4 contains a Clustal W (Thompson et al., 1994) alignment of bactridines 1 and 2, together with ardiscretin and TsNTx. Although the 4 peptides have diverse degrees of sequence identity (consensus of the other 3 peptides with bactridine 1 are underlined in the table), the sequence identities between ardiscretin and bactridine 1 are striking (wave underlined residues in the table). From the AA sequence of P. Dı́az et al. / Toxicon 54 (2009) 802–817 Fig. 4. Effects of bactridines 1, 2 and 4 on the Na-CoroNa red (1 mM) complex concentration inside and outside of Y. enterocolitica bacteria and the antagonims by amiloride (10 mM) and mibefradil (25 mM) on the bactridines’ effect. The effect of each bactridine per se, inside the bacteria or in their neighborhood are represented as -. Ordinate is integrated optical density inside the bacteria and in the immediate external vicinity of the bacterial membrane, measured as indicated in Methods. Bactridines 1, 2 and 4 produced a large flow of Na-CoroNa red complex bacterial cytoplasm to the neighbouring external medium. Both amiloride (:) and mibefradil (C) were able to block completely the outflow Na-CoroNa red complex. The symbols are medians, bars on top of symbols are the 95% confidence intervals of the medians of 20 cells. The data are presented as median and its 95% confidence interval, 10 bacteria per point. 811 812 P. Dı́az et al. / Toxicon 54 (2009) 802–817 Fig. 5. Effects of bactridines 3, 5 and 6 on the Na-CoroNa red (1 mM) complex concentration inside and outside of Y. enterocolitica bacteria and the antagonims by amiloride (10 mM) and mibefradil (25 mM) on the bactridines’ effect. The effect of each bactridine per se, inside the bacteria or in their neighborhood are represented as -. While bactridine 3 produced a large flow of Na-CoroNa red complex bacterial cytoplasm to the neighbouring external medium, bactridines 5 and 6, in turn, were less potent. Amiloride (:) was able to block completely the outflow Na-CoroNa red complex induced by the three bactridines. Mibefradil (C), in turn antagonised completely the effect of bactridine 5, but antagonised partially the effect of bactridine 3, and did not antagonised the effect of bactridine 6. The data are presented as median and its 95% confidence interval, 10 bacteria per point. Other details are as in Fig. 4. P. Dı́az et al. / Toxicon 54 (2009) 802–817 813 Fig. 6. Effect of bactridines 1 and 2 on ionic outflow from Y. enterocolitica in 0.15 M choline chloride. Top panels: effects of bactridines 1 and 2 on the Na-CoroNa red (1 mM) complex concentration inside (left panel) and outside (right panel) Y. enterocolitica bacteria. Middle panels: similar to the top panels but from bacteria loaded with the Ca2þ selective fluorescent dye Fluo 4AM (10 mM). Bottom panels: similar to the top panels but from bacteria loaded with the Kþ selective fluorescent dye PBFI (20 mM). In all panels, C: are the control, that is, bacteria where an aliquot of just the choline chloride solution was added in the microscope; ;: indicate results obtained in bacteria exposed to bactridine 1, and, :: indicate results obtained in bacteria exposed to bactridine 2. Both bactridines induced a large increase in Na-CoronNa red in the vicinity of the bacteria, no changes in Ca-Fluo 4AM- or K-PBFI-complex were detected neither inside nor outside the bacteria. The data are presented as median and its 95% confidence interval, the number of data per point in 45 experimental points was 23 (17, 29) bacteria and ranged from 7 to 45 bacteria, the exact number in each condition is provided as supplementary material. Other details are as in Fig. 3 and in the text of the communication. bactridine 1 and 2 it was possible to infer that they are positively charged (Gasteiger et al., 2005) like all known antibacterial peptides from venomous arthropod’s haemolymph or venoms (Tables 1–3, Kuhn-Nentwig, 2003). Theoretical grounds have been established to determine when the result of in silico modeling predict a tertiary structure accurately, and to define, the so called, ‘‘safe homology modeling zone’’. In this zone similar sequences adopt practically identical structures, and distantly related sequences still fold into similar structures. Some of the theoretical grounds were set by Chothia and Lesk (1986), Sander and Schneider (1991) and Rost (1999), and are 814 P. Dı́az et al. / Toxicon 54 (2009) 802–817 Table 3 Percent haemolysis induced by bactridines measured from haemoglobin absorbance in plasma. Concentrations 90 mM 180 mM Bactridine Bactridine Bactridine Bactridine Bactridine Bactridine Triton 1% 0.2 [0.2, 0.3]% 7.6 [7.4, 7.8]% (0.1 [0.0, 0.3])% (4.3 [4.2, 4.4])% (0.9 [0.8, 0.9])% (3.1 [3.0, 3.1])% 100% 1.0 [0.9, 1.1]% 21.0 [21.0, 21.0]% (0.5 [0.4, 0.6])% (26.1 [25.6, 26.7])% (0.9 [0.8, 0.9])% (12.1 [11.7, 12.5])% 100% 1 2 3 4 5 6 The concentration corresponding to percent haemolysis without parentheses were calculated setting the absorbance per unit path length (a0 ) as 3.242 for bactridine 1 and as 3.720 for bactridine 2. Values between parentheses were calculated assuming that a0 was 3.481 (Average of the a0 values of bactridines 1 and 2). All percentages are the mean of two assays, range between brackets, all data rounded to the first decimal. See Methods and Results sections for other details. summarized by Krieger et al. (2003). In modeling bactridines 1 and 2 the similarities with the template peptide used, guarantee that our homology modeling is within the so called ‘‘safe homology modeling zone’’, which exist for peptides of z60 AA when the identity between the target sequence and the template used for homology modeling is 60%. The three dimensional models obtained by means of homology modeling strongly suggest that bactridines 1 and 2 shown in Fig. 3 are globular proteins which fold according to the cysteine a–b-motif. These peptides have an a-helix segment and a conserved b-sheet structure maintained by disulfide bonds. The concentrations used for toxicological and physiological experiments for ardiscretin in D’Suze et al. (2004) (all done prior to sequencing the peptides) were calculated assuming that a0 ¼ 1, we now re-calculated the values of a0 for ardiscretin (for absorbance at 280 nm) in the same manner done here for the bactridines and got 4.458 assuming that all Cys appear as half cystines. Thus, using a0 ¼ 4.458, in D’Suze et al. (2004) ardiscretin: 0.7 mg/g were non-toxic to mice, 0.9–1.6 mg/g killed U. rapax crabs, 12–22 mg/g killed crickets (Achaeta sp.) and 33–300 mg/g killed triatomines (Rhodnius prolixus). In this communication bactridine 1, 1.3 mg/g did not kill mice, a dose approximately equal to twice the dose of ardiscretin used by D’Suze et al. (2004); but bactridine 1 at a dose as low as 0.9 mg/g induced sialorrhoea in crabs, when the dose was raised to 1.8 mg/g the sialorrhoea was more intense and 43% crab died. Cockroaches (P. americana, z1 g bw) injected through a leg’s first joint with z22 mg/g of bactridine 1, died; this dose is close to the one killing crickets, and close to the lower bound of the dose killing triatomines in D’Suze et al. (2004). These results indicate that ardiscretin and bactridine 1 are approximately equally toxic to vertebrates and invertebrates. Bactridine 2 had a 98% sequence identity with the first 64 AA of Tz1 of T. zulianus and Td4 of T. discrepans (Borges et al., 2004). The sequences differ in the Asn34 in bactridine 2 and an Asp34 in Tz1 and Td4. Bactridine 5 had molecular mass of 7101 Da, very close to ardiscretin which has a molecular mass of 7103 Da (Batista et al., 2006); yet, similar masses do not mean equal sequences. Bactridine 5 was found to affect mice at z1 mg/ g, and had no effect on cockroaches even using z20 mg/g roach, making it 20 times more potent in the mammal than in the insect; together with bactridines 2, 3 and 4, bactridine 5 was completely ineffective in crabs. Ardiscretin, 0.7 mg/g animal, had no effect in mice but at roughly the same dose killed crabs, and killed crickets at approximately the same dose bactridine 5 did not kill cockroaches. Two chemically different molecules may have the same effect, but if the effect of two molecules, studied in the same animals (mice and crabs) under similar conditions and at similar concentrations differ, their structures must be different. This would appear to be the case for bactridine 5 and ardiscretin. Bactridines were effective against several Gram positive and Gram negative bacteria at concentrations ranging 20–77 mM depending on the bacteria and bactridine tested (Table 2). This range is in very reasonable agreement with the MICs of other peptides isolated from arthropods haemolymph and venom against Gram positive bacteria (3.3 to 150 mM, Kuhn-Nentwig 2003) or against Gram negative bacteria (0.7–17.1 mM, Kuhn-Nentwig 2003). Since Table 4 Clustal W 1.18 alignment of bactridines 1 and 2, ardiscretin and T. serrulatus non-toxic peptide TsNTx P. Dı́az et al. / Toxicon 54 (2009) 802–817 bactridines were detected in different venom lots, they seem to be integral components of T. discrepans scorpion venom; this suggests that bactridines play a significant physiological role in this scorpion. Interestingly, scorpions, at least in captivity, spray their venom over their bodies in a manner suggesting a cleansing or protective practice (authors’ unpublished observation). Our experiments with CoroNaÔ red-loaded bacteria in both distilled water (bactridines 1–6) or 0.15 M choline chloride (bactridines 1 and 2, other bactridines nor tested), show that a sodium efflux occurs when Y. enterocolitica is exposed to bactridines. While the effect of bactridines in distilled water was evident as a decrease in IOD inside and an increase in the neighborhood of the bacteria, in 0.15 M choline chloride only an increase of Naþ in the vicinity of the bacteria was evident. If an increase in fluorescence occurs in the vicinity of the bacteria observed in a two dimensional picture, the increase occurring above and bellow the bacteria could mask a decrease in IOD within the bacteria. The increases in IOD in the neighborhood of the bacteria exposed to bactridines 1 or 2 in 0.15 M choline chloride, were approximately as large as the extracellular increases observed in distilled water. One interpretation of this could be that the decreases within the bacteria in 0.15 M choline chloride were somewhat smaller than in distilled water, and that the Na–CoroNa red complex was more fluorescent in choline chloride. One point to consider is that the effect of bactridines on microbes using CoroNaÔ Red were observed at 1 mM and the MIC observed are 1 or 2 orders of magnitude larger. This apparent incongruency may be understood considering that the fluorescent method is very sensible, as most modern fluorescent dyes, CoroNaÔ Red has a high quantum efficiency. So it indicates sodium ion concentration changes at low toxin concentrations. Actually, bactridines’s effects on microbes were studied in culture medium, which contains sodium and many proteins, so in such media bacteria are loaded with sodium not depleted as in the CoroNaÔ Red experiments presented here; but loading or depleting of sodium is due to sodium channel opening and the choice of Naþ concentration in the extracellular space. Also, although most of the bactridine molecules are probably free to move and interact with bacteria in the extracellular medium used for fluorescence experiments, this might not be the case in culture media where the peptides could partially form complexes with the peptides and other organic molecules which abound in the culture media. We also studied possible changes in intra- and extracellular Ca2þ and Kþ concentrations using Fluo 4AM and PBFI, respectively. As seen in the middle and bottom panels of Fig. 6, the fluorescence on neither Fluo 4AM nor PBFI were distinct from their controls, when observed after 30 s to 1 min after adding the bactridines. We thus, found no evidence that these peptides alter either intracellular Kþ or Ca2þ concentrations. These experiments showed that IOD within the bacteria and in its vicinity may be transiently affected by the perturbation induced when an aliquot of anything is added to the bacteria in the microscope; this was evident for the control run (C in the middle left panel of Fig. 6) and for the application of bactridine 2 (: in the middle left, panel as well as in the middle right and lower 815 right panels of the same figure). Also, the fluorescence of both Fluo 4AM and PBFI were evidently less intense than the Na–CoroNa Red complex, their fluorescence measurements were subject of greater uncertainty. Bactridine 1 induced Na–CoroNa red complex outflow was blocked by 10 mM amiloride and by 25 mM mibefradil. Amiloride is known to block amphibian skin and kidney non voltage gated Naþ channels (Garty and Palmer, 1997) and it is a potent inhibitor of the Naþ driven flagellar motion of Bacillus firmus (Atsumi et al., 1990) and blocks the Naþ channel dependent uptake in the thermophilic Bacillus strain TA2.A1 (Peddie et al., 1999); this suggests that amiloride also blocks voltage gated Naþ channels (Nav) in bacteria. Mibefradil is a T- and L-type Ca channel blocker (Abermethy, 1997), which also blocks the NaChBac Nav from Bacillus halodurans (Yue et al., 2002). These evidences suggest that bactridine affects bacteria by opening sodium channels. Tetrodotoxin (30 mM), a highly specific Nav blocker in nerve and muscle in nanomolar concentrations (Narahashi et al., 1964), has no effect on NaChBac channels (Ren et al., 2001) and did not antagonize bactridineinduced sodium outflow. Our experiments clearly show that bacteria exposed to a very high dose of TTX are still depleted from Naþ by bactridines. NaChBac channels are very sensitive to L-type Ca2þ channel blockers such as nifedipine (IC50 1 mM) and nimodipine (IC50 2 mM), but relatively insensitive to mibefradil (IC50 22 mM, Ren et al., 2001). It is an interesting finding that the sodium outflow induced by bactridines 1, 2, 4 and 6 was completely blocked by 25 mM mibefradil, while the same concentration of mibefradil only blocked partially, if at all, the outflow induced by bactridines 3 and 5. The different sensitivities to mibefradil may indicate 2 isoforms of Nav in Y. enterocolitica; one of them with pharmacological properties similar to NaChBac, and another one more sensitive to mibefradil. The existence of Nav isochannels is not new, it was first demonstrated in squid nerves by Sevcik (1976, 1982). Yet, since our studies do not definitively rule out the effect of bactridines in sodium translocation pathways different from ion channels, additional investigation is required to settle this point. Out of the 6 bactridines isolated, bactridine 1 was neither haemolytic nor otherwise toxic to mice; this lack of mice toxicity suggests that bactridine 1 could function as a model of potentially useful antibiotics for use in mammalians. This is remarkable since all antibiotic peptides described to date, although often more potent as antibacterials than bactridine 1, are either haemolytic or otherwise toxic to mammalians, citing Kuhn-Nentwig (2003): ‘‘Despite their excellent antimicrobial activities, we have to realize that antimicrobial and cytolytic peptides from arthropod venom are at present not well suited for new antibiotic drugs since many of them do not sufficiently discriminate between microorganisms (pathogens) and erythrocytes (eukaryotic cells). Up to now, numerous efforts have attempted to elucidate the structural basis of their broad cytolytic activity, but the underlying mechanisms are still not well understood’’ Thus, bactridine 1 points in the direction of the structural characteristics of peptides that have differential 816 P. Dı́az et al. / Toxicon 54 (2009) 802–817 effects on pathogens and eukaryotic cells. Our work also demonstrates that a new way to design antibiotics may be related to designing drugs, peptidic or not, targeted at bacterial sodium channels, which differ enough from channels on eukaryotic cells to permit bacteria death only, and that drug killing bacteria via sodium permeability increase do not need to produce cytolysis of eukaryotic cells. It must be pointed out, however, that finding an antibiotic peptide in a scorpion venom, or any other of its fluids for that matter, constitutes no proof per se of it usefulness as a possible novel antibiotic; as with any other natural product, more research is needed to validate the clinical usefulness of bactridine 1 as a medically significant drug. Acknowledgements The authors are indebted to the people of San Antonio de Los Altos and their Fire Department for the supply of the scorpions. We greatly appreciated the technical assistance of Lic. Moisés Sandoval. Dr. Carlo Caputo kindly provided mibefradil and M.Sc. Hildemaro López provided amiloride. The Graphic Design Department at IVIC produced Fig. 1. This research was partially supported by the Venezuelan FONACIT grant No. S1-2001000908, FONACIT grant to IVIC’s proyect 416, and by Laboratorios Silanes/Instituto Bioclón, Mexico. Conflict of interest The authors declare there are no conflicts of interest. Appendix. Supplementary data Supplementary data associated with this article can be found in the online version, at doi:10.1016/j.toxicon.2009. 06.014. References Abermethy, D.R., 1997. Pharmacologic and pharmacokinetic profile of mibefradil a T- and L-type calcium channel antagonist. Am. J. Cardiol. 80, 4C–11C. Ahn, A.H., Basbaum, A.I., 2006. Tissue injury regulates serotonin 1d receptor expression: implications for the control of migraine and inflammatory pain. J. Neurosci. 26, 8332–8338. Altschul, S.F., Thomas, L., Madden, T.L., Schäffer, A.A., Zhang, J., Zhang, Z., Miller, W., Lipman, D.J., 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25, 3389–3402. Arnold, K., Bordoli, L., Kopp, J., Schwede, T., 2006. The SWISS-MODEL workspace: a web-based environment for protein structure homology modelling. Bioinformatics 22, 195–201. Atsumi, T., Sugiyama, S., Cragoe Jr., E.J., Imae, Y., 1990. Specific inhibition of the Naþ-driven flagellar motors of alkalophilic Bacillus strains by the amiloride analog phenamil. J. Bacteriol. 172, 1634–1639. Baron, S., Caplanusi, A., Van de Ven, M., Radu, M., Despa, S., Lambrichts, I., Ameloot, M., Steels, P., Smets, I., 2005. Role of mitochondrial Naþ concentration, measured by CoroNa red, in the protection of metabolically inhibited MDCK cells. J. Am. Soc. Nephrol. 16, 3490–3497. Batista, C.V., D’Suze, G., Gòmez-Lagunas, F., Zamudio, F.Z., Encarnación, S., Sevcik, C., Possani, L.D., 2006. Proteomic analysis of Tityus discrepans scorpion venom and amino acid sequence of novel toxins. Proteomics. 12, 3718–3727. Bernardinelli, Y., Azarias, G., Chatton, J.-Y., 2006. In situ fluorescence imaging of glutamate-evoked mitochondrial Naþ responses in Astrocytes. Glia 54, 460–470. Boman, H.G., 2003. Antibacterial peptides: basic facts and emerging concepts. J. Intern. Med. 254, 197–215. Borges, A., Alfonzo, M.J., Garcı́a, C.C., Winand, N.J., Leip-old, E., Heinemann, S.H., 2004. Isolation, molecular cloning and functional characterization of a novel b-toxin from the Venezuelan scorpion, Tityus zulianus. Toxicon 43, 671–684. Bulet, P., Stöcklin, R., Menin, L., 2004. Anti-microbial peptides: from invertebrates to vertebrates. Immunol. Rev. 198, 169–184. Ceroni, A., Passerini, A., Vullo, A., Frasconi, P., 2006. DI-SULFIND: a disulfide bonding state and cysteine connectivity prediction server. Nucleic Acids Res. 34 (Web Server issue), W177–W181. Chang, C.T., Wu, C.S., Yang, J.T., 1978. Circular dichroic analysis of protein conformation: inclusion of the b-turns. Anal. Biochem. 9, 13–31. Chothia, C., Lesk, A.M., 1986. The relation between the divergence of sequence and structure in proteins. EMBO J. 5, 823–836. Cociancich, S., Goyffon, M., Bontems, F., Bulet, P., Bouet, F., Menez, A., Hoffman, J., 1993. Purification and characterization of a scorpion defensin, a 4 kDa antibacterial peptide presenting structural similarities with insect defensins and scorpion toxins. Biochem. Biophys. Res. Commun. 194, 17–22. Corzo, G., Adachi-Akahane, S., Nagao, T., Kusui, Y., Nakajima, T., 2001. Novel peptides from assassin bugs (Hemiptera: Reduviidae): isolation, chemical and biological characterization. FEBS Lett. 499, 256–261. Conde, R., Zamudio, F.Z., Rodrı́guez, M.H., Possani, L.D., 2000. Scorpine, an anti-malaria and anti-bacterial agent purified from scorpion venom. FEBS Lett. 471, 165–168. Dai, L., Corzo, G., Naoki, H., Andriantsiferana, M., Nakajima, T., 2002. Purification, structure-function analysis, and molecular characterization of novel linear peptides from scorpion Opisthacanthus madagascariensis. Biochem. Biophys. Res. Commun. 293, 1514–1522. Dı́az-Bello, P., D’Suze, G., Salazar, V., Sevcik C., Fox, J., 2007. Antibacterial activity of six novel peptides. A fluorescent probe study of microbial membrane Naþ permeability changes. In: 9th Panamerican Conference on Animal, Plant & Microbial Toxins. 21–25 October, Juriquilla, México. D’Suze, G., Sevcik, C., Corona, M., Zamudio, F.Z., Batista, C.V.F., Coronas, F.I., Possani, L.D., 2004. Ardiscretin a novel arthropod-selective toxin fron Tityus discrepans scorpion venom. Toxicon 43, 263–272. D’Suze, G., Sevcik, C., Ramos, M., 1995. Presence of curarizing polypeptides and a pancreatitis-inducing fraction without muscarinic effects in the venom of the Venezuelan scorpion Tityus discrepans (Karsch). Toxicon 33, 333–345. Ehret-Sabatier, L., Loew, D., Goyffon, M., Fehlbaum, P., Hoffmann, J.A., van Dorsselaer, A., Bulet, P., 1996. Characterization of novel cysteine-rich antimicrobial peptides from scorpion blood. J. Biol. Chem. 271, 29537–29544. Epand, R., Vogel, H., 1999. Diversity of antimicrobial peptides and their mechanisms of action. Biochim. Biophys. Acta 1462, 11–28. Garty, H., Palmer, L., 1997. Epithelial sodium channels: function, structure and regulation. Physiol. Rev. 77, 359–396. Gasteiger, E., Hoogland, C., Gattiker, A., Duvaud, S., Wilk-ins, M.R., Appel, R.D., Bairoch, A., 2005. Protein Identification and Analysis Tools on the ExPASy Server. In: Walker, John M. (Ed.), The Proteomics Protocols Handbook. Humana Press, pp. 571–607. Gee, K.R., Brown, K.A., Chen, W.N., Bishop-Stewart, J., Gray, D., Johnson, I., 2000. Chemical and physiological characterization of Fluo-4 Ca2þindicator dyes. Cell Calcium 27, 97–106. Giangaspero, A., Sandri, L., Tossi, A., 2001. Amphipathic alpha helical antimicrobial peptides. Eur. J. Biochem. 268, 5589–5600. Guatimosim, S.C., Prado, V.F., Diniz, C.R., Chavez-Olorte-gui, C., Kalapothakis, E., 1999. Molecular cloning and genomic analysis of TsNTxp: an immunogenic protein from Tityus serrulatus scorpion venom. Toxicon 37, 507–517. Guex, N., Peitsch, M.C., 1997. SWISS-MODEL and the Swiss-PdbViewer: an environment for comparative protein modelling. Electrophoresis 18, 2714–2723. Hetru, Ch., Letellier, L., Oren, Z., Hoffmann, J.A., Shai, Y., 2000. Androctonin, a hydrophilic disulphide-bridged non-haemolytic anti-microbial peptide: a plausible mode of action. Biochem. J. 345, 653–664. Iwamoto, M., Allen, R.D., 2004. Uptake and rapid transfer of fluorescent ceramide analogues to acidosomes (late endosomes) in paramecium. J. Histochem. Cytochem. 52, 555–565. Ješek, P., Mahdi, F., Garlid, K.D., 1990. Reconstitution of the beef heart and rat liver mitochondrial Kþ/Hþ (Naþ/Hþ) antiporter. Quantitation of Kþ transport with the novel fluorescent probe, PBFI. J. Biol. Chem. 265, 10522–10526. Kondejewski, L.H., Farmer, S.W., Wishart, D.S., Kay, C.M., Hancock, R.E., Hodges, R.S., 1996. Modulation of structure and antibacterial and hemolytic activity by ring size in cyclic gramicidin S analogs. J. Biol. Chem. 271, 25261–25268. P. Dı́az et al. / Toxicon 54 (2009) 802–817 Krieger, E., Darden, T., Nabuurs, S.B., Finkelstein, A., Vri-end, G., 2004. Making optimal use of empirical energy functions: force field parametrization in crystal space. Proteins 57, 678–683. Krieger, E., Nabuurs, S.B., Vriend, G., 2003. Introduction to homology modeling. Methods Biochem. Anal. 44, 509–523. Kuhn-Nentwig, L., 2003. Antimicrobial and cytolytic peptides of venomous arthropods. Cell Mol. Life Sci. 60, 2651–2668. Lehrer, R.I., Ganz, T., Selsted, M.E., 1991. Defensins: endogenous antibiotic peptides of animal cells. Cell 64, 229–230. Matsuzaki, K., Sugishita, K., Miyajima, K., 1999. Interactions of an antimicrobial peptide, magainin 2, with lipopolysaccharide-containing liposomes as a model for outer membranes of gram-negative bacteria. FEBS Lett. 449, 221–224. Moerman, L., Bosteels, S., Noppe, W., Willems, J., Clynen, E., Schoofs, L., Thevissen, K., Tytgat, J., Van Eldere, J., Van Der Walt, J., Verdonck, F., 2002. Antibacterial and antifungal properties of a-helical, cationic peptides in the venom of scorpions from southern Africa. Eur. J. Biochem. 269, 4799–4810. Narahashi, T., Moore, J.W., Scott, W., 1964. Tetrodotoxin blockage of sodium conductance increase in lobster giant axons. J. Gen. Physiol. 47, 965–974. Ovchinnikova, T.V., Balandin, S.V., Aleshina, G.M., Tagaev, A.A., Leonova, Y.F., Krasnodembsky, E.D., Men’shenin, A.V., Kokryakov, V.N., 2006. Aurelin, a novel antimicrobial peptide from jellyfish Aurelia aurita with structural features of defensins and channel-blocking toxins. Biochem. Biophys. Res. Commun. 348, 514–523. Peddie, C.J., Cook, G.M., Morgan, H.W., 1999. Sodium-dependent glutamate uptake by an alkaliphilic, thermophilic Bacillus strain, TA2.A1. J. Bacteriol. 181, 3172–3177. Pinheiro, C.B., Marangoni, S., Toyama, M.H., Polikarpov, I., 2003. Structural analysis of Tityus serrulatus Ts1 neurotoxin at atomic resolution: insights into interactions with Naþ channels. Acta Crystallogr., Sect. D 59, 405–415. Ren, D., Navarro, B., Xu, H., Hue, L., Shi, Q., Clapham, D.E., 2001. A prokaryotic voltage-gated sodium channel. Science 294, 2372–2375. Rodrı́guez de la Vega, R.C., Garcı́a, B.I., D’Ambrosio, C., Diego-Garcı́a, E., Scaloni, A., Possani, L.D., 2004. Antimicrobial peptide induction in the haemolymph of the Mexican scorpion Centruroides limpidus limpidus in response to septic injury. Cell Mol. Life Sci. 61, 1507–1519. Rost, B., 1999. Twilight zone of protein sequence alignments. Protein Eng. 12, 85–94. Sander, C., Schneider, R., 1991. Database of homology-derived protein structures and the structural meaning of sequence alignment. Proteins 9, 56–68. 817 Schwede, T., Kopp, J., Guex, N., Peitsch, M.C., 2003. SWISS-MODEL: an automated protein homology-modeling server. Nucleic Acids Res. 31, 3381–3385. Senko, M.W., Beu, S.C., McLafferty, F.W., 1995. Determination of monoisotopic masses and ion populations for large biomolecules from resolved isotopic distributions. J. Am. Soc. Mass Spectrom. 6, 229–233. Sevcik, C., 1976. Binding of tetrodotoxin to squid nerve fibers. Two kinds of receptors? J. Gen. Physiol. 68, 95–103. Sevcik, C., 1982. Temperature dependence of tetrodotoxin effect in squid giant axons. J. Physiol. 325, 187–194. Steiner, H., Hultmark, D., Engstrom, A., Bennich, H., Boman, H.G., 1981. Sequence and specificity of two antibacterial proteins involved in insect immunity. Nature 292, 246–248. Thompson, J.D., Higgins, D.G., Gibson, T.J., 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, positions-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22, 4673–4680. Torres-Larios, A., Gurrola, G.B., Zamudio, F.Z., Possani, L.D., 2000. Hadrurin, a new antimicrobial peptide from the venom of the scorpion Hadrurus aztecus. Eur. J. Biochem. 267, 5023–5031. Travis, S.M., Anderson, N.N., Forsyth, W.R., Espiritu, C., Conway, B.D., Greenberg Jr., E.P., McCray, P.B., Lehrer, R.I., Welsh, M.J., Tack, B.F., 2000. Bactericidal activity of mammalian cathelicidin-derived peptides. Infect. Immun. 68, 2748–2755. Uawonggul, N., Thammasirirak, S., Chaveerach, A., Arkaravichien, T., Bunyatratchata, W., Ruangjirachuporn, W., Jearranaiprepame, P., Nakamura, T., Matsuda, M., Kobayashi, M., Hattori, S., Daduang, S., 2007. Purification and characterization of Heteroscorpine-1 (HS-1) toxin from Heterometrus laoticus scorpion venom. Toxicon 49, 19–29. Wishart, D.S., Bigam, C.G., Yao, J., Abildgaard, F., Dyson, H.J., 1995. 1H, 13C and 15N chemical shift referencing in biomolecular NMR. J. Biomol. NMR 6, 135–140. Yue, L., Navarro, B., Ren, D., Ramos, A., Clapham, D., 2002. The cation selectivity filter of the bacterial sodium channel, NaChBac. J. Gen. Physiol. 120, 845–853. Zasloff, M., 1987. Magainins, a class of antimicrobial peptides from Xenopus skin: isolation, characterization of two active forms, and partial cDNA sequence of a precursor. Proc. Natl. Acad. Sci. U.S.A. 84, 5449–5453. Zhu, S., Li, W., Jiang, D.H., Zeng, X.-C., 2000. Evidence for the existence of insect defensin-like peptide in scorpion venom. IUBMB Life 50, 57–61.