Lycotoxins, Antimicrobial Peptides from Venom of the Wolf SpiderLycosa carolinensis *

Two peptide toxins with antimicrobial activity, lycotoxins I and II, were identified from venom of the wolf spiderLycosa carolinensis (Araneae: Lycosidae) by virtue of their abilities to reduce ion and voltage gradients across membranes. Both peptides were purified to homogeneity by reversed-phase liquid chromatography and determined to have the following primary structures by Edman microsequencing: IWLTALKFLGKHAAKHLAKQQLSKL-NH2 for lycotoxin I and KIKWFKTMKSIAKFIAKEQMKKHLGGE-OH for lycotoxin II. The predicted secondary structures of the lycotoxins display amphipathic α-helix character typical of antimicrobial pore-forming peptides. Antimicrobial assays showed that both lycotoxins potently inhibit the growth of bacteria (Escherichia coli) and yeast (Candida glabrata) at micromolar concentrations. To verify its hypothesized pore-forming activity, lycotoxin I was synthesized and shown to promote efflux of Ca2+ from synaptosomes, to cause hemolysis of erythrocytes, and to dissipate voltage gradients across muscle membrane. The lycotoxins may play a dual role in spider-prey interaction, functioning both in the prey capture strategy as well as to protect the spider from potentially infectious organisms arising from prey ingestion. Spider venoms may represent a potentially new source of novel antimicrobial agents with important medical implications.

Most spiders are voracious predators, employing paralytic venoms to immobilize or kill their prey. The majority of known spider venom toxins attack the nervous system, causing disruption of impulse conduction and/or synaptic transmission through actions on ion channels (1-3) and exocytosis proteins (4,5). Alternatively, many spider venoms cause cell membrane disruption and consequent tissue necrosis through enzymatic actions (6 -8).
The wolf spider Lycosa carolinensis (Araneae: Lycosidae) is an omnivorous hunting spider with a widespread distribution in the United States. The spider hides underground during the day, but leaves its burrow during the night in search of prey. The effects of wolf spider bites on humans are not considered to be severe (9), and relatively little is known with regard to the chemical composition of L. carolinensis venom. Previous studies involving related species indicate the presence of neuroac-tive constituents (10).
Here we characterize two peptide toxins from L. carolinensis venom that are both neuroactive and antimicrobial. Named here as "lycotoxins," they appear from their amphipathic character and physiological actions to function as pore formers to increase membrane permeability and to effect lysis of both prokaryotic and eukaryotic cells.

EXPERIMENTAL PROCEDURES
Toxin Purification-Wolf spiders (Lycosa carolinensis Araneae: Lycosidae) were collected from Yavapai County, AZ and held in individual containers in the Department of Entomology, University of California, Riverside, CA. Whole venom was obtained through an electrical milking technique and held at Ϫ80°C until processed. Prior to fractionation by reversed-phase liquid chromatography, venom was dissolved in 0.1% aqueous trifluoroacetic acid and fractionated with a Brownlee widepore C 8 column (4.6 ϫ 150 mm) using a linear gradient of aqueous acetonitrile (CH 3 CN) in constant 0.1% trifluoroacetic acid at a flow rate of 1.0 ml/min. Subsequent purification of biologically active peaks involved variation of organic modifier (acetonitrile or n-propyl alcohol), ion-pairing agent (trifluoroacetic acid or heptafluorobutyric acid), or column (Brownlee C 8 or Vydac C 4 ).
Synaptosomal 45 Ca 2ϩ Measurements-In most experiments, 45 l of resuspended synaptosomes (ϳ180 g of membrane proteins) was incubated with 5 l of test fraction or deionized water (control) for a specified period of time. To this was added 0.8 M 45 Ca 2ϩ in 50 l of either "high potassium" saline (same as calcium-free saline but containing 137 mM KCl and 1.0 mM CaCl 2 ) or "low potassium" saline (similar to calcium-free saline but containing 1.0 mM CaCl 2 ). Following a 3-s incubation, the reaction was stopped by addition of 200 l of termination solution (30 mM EGTA, 120 mM NaCl, and 5 mM KCl, adjusted to pH 7.6 with Tris base). Extrasynaptosomal 45 Ca 2ϩ was removed by rapid filtration with washing buffer (145 mM NaCl, 5 mM KCl, 1 mM CaCl 2 , 1.4 mM MgCl 2 , and 20 mM HEPES at pH 7.4) using a Skatron cell harvester. Filters were immersed in 4 ml of Beckman Ready Safe scintillation fluid and subjected to scintillation counting.
Mass Spectrometry, Amino Acid Composition, and Sequencing Analyses-All analyses were performed at the Biotechnology Instrumentation Facility, University of California, Riverside. Matrix-assisted laser desorption time-of-flight mass analyses were performed with a Finnigan Lasermat instrument. Amino acid composition and sequencing analyses were performed as described previously (12).
Peptide Synthesis-Synthetic lycotoxin I was prepared by the Sussex Center for Neuroscience, University of Sussex, Sussex, United Kingdom, on an Applied Biosystems Model 432A Synergy automated peptide synthesizer. The peptide was prepared by solid-phase synthesis using Fmoc (N-(9-fluorenyl)methoxycarbonyl) N-terminal protection of amino acids, and amino acids were activated to form an active ester by 2-(1-H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate. Peptide cleavage from the resin and side chain deprotection were achieved using trifluoroacetic acid and the scavengers thioenisole and ethanedithiol (100 l each). Cleavage was conducted for 3 h at room temperature, followed by extraction of the peptide using methyl-t-butyl ether to separate scavengers and reaction by-products. The peptide was lyophilized and purified by preparative RPLC 1 using an Aquapore Octyl Prep 10 cartridge column (10 ϫ 250 mm) at 5 ml/min with a gradient of 0 -60% solvent B (solvent A ϭ 5% CH 3 CN and 0.1% trifluoroacetic acid; solvent B ϭ 60% CH 3 CN and 0.08% trifluoroacetic acid).
Antimicrobial Assays-Magainin B (13), a synthetic peptide analog related to magainin 2 isolated from Xenopus laevis (14), was a generous gift from Dr. Hao-Chia Chen (National Institutes of Health, Bethesda, MD  (15) to exponential phase with an A 600 of 0.8, representing 10 9 colony-forming units/ml. Bacteria were added to warm (ϳ45°C) 0.7% agar in LB broth at a ratio of 1 l of bacteria/ml of medium. Then, 8 or 35 ml of the bacteriaagarose/LB mixture was poured over the bottom of a 1.5% agarose/LB plate in a 100 ϫ 20-or 150 ϫ 20-mm dish. 5 l of magainin B or lycotoxins of various concentrations was then applied to the solidified plate surface as discrete drops. After incubation at 37°C overnight, the effects of magainin B and lycotoxins were recorded as the clear spots in the bacterial lawn. The antimicrobial activities of lycotoxins and magainin B on E. coli D31 in liquid culture were determined in 96-well plates in a final volume of 100 l as follows. 10 l of different concentrations of peptides was added to 90 l of LB broth containing the inoculate of E. coli D31 adjusted to 10 5 to 10 6 colony-forming units/ml. After 7 h of incubation at 37°C, inhibition of growth was determined by measuring the absorbance at 570 nm. Each minimal inhibitory concentration was determined from two independent experiments performed in duplicate.
C. albicans and C. glabrata were grown overnight in YTD medium (1% yeast extract, 2% Tryptone, and 2% glucose) at 37°C, giving an A 600 of ϳ0.9 -1.0 (representing 10 7 colony-forming units/ml). 100 l of the yeast was diluted in 10 ml of sterilized water. 500 l of diluted yeast FIG. 1. Isolation of lycotoxins from venom of the wolf spider L. carolinensis. A, whole venom (25 l) was fractionated on a Brownlee C 8 column (300 Å; 4.6 ϫ 250 mm) using a linear gradient of acetonitrile/ water (0.5%/min) in constant 0.1% trifluoroacetic acid. Elution of lycotoxins I and II is indicated by arrows. B, shown is the elution profile of lycotoxin I using a linear gradient of n-propyl alcohol/water (0.33%/min) in constant 0.5% trifluoroacetic acid (same column as described for A). C, shown is the elution profile of 7.5 nmol of native lycotoxin I using a Vydac C 4 column (300 Å; 4.6 ϫ 250 mm) and an acetonitrile/water gradient (0.25%/min) in constant 0.1% heptafluorobutyric acid. D, shown is the coelution of native lycotoxin I (4 nmol) and synthetic lycotoxin I (3 nmol) using the same conditions as described for C. mAU, milli-absorbance units.
a The values in parentheses are from Edman sequencing. b The presence of Trp was based on UV absorption spectrum.
was plated evenly on each 150 ϫ 20-mm agarose plate of YTD medium. 5 l of magainin B or lycotoxins of various concentrations was then applied to the solidified plate as discrete drops. After incubation at 37°C overnight, the effects of magainin B and lycotoxins were recorded as the clear spots in the yeast lawn. Hemolysis Assay-Hemolytic activity of both lycotoxin I and magainin B was assayed with heparinized red blood cells from rabbit (Oryctolagus cuniculus) rinsed three times in PBS (50 mM NaH 2 PO 4 and 150 mM NaCl, pH 7.0) by centrifugation for 5 min at 4310 ϫ g. Red blood cells were then incubated at room temperature for 1 h in deionized water (positive control), in PBS (blank), or with synthetic lycotoxin I or magainin B at various concentrations (6.25-200 M) in PBS. The samples were centrifuged at 11,640 ϫ g for 3 min. The supernatant was separated from the pellet, and its absorbance was measured at 570 nm. Zero hemolysis (blank) and 100% hemolysis controls were determined following centrifugation of the centrifugates of red blood cells suspended in PBS and H 2 O, respectively. Neuromuscular Pharmacology-Lycotoxin I was assayed on the body wall muscles of prepupal house flies (Musca domestica, NAIDM strain) using the intracellular electrophysiological recording methods described previously (16).

RESULTS
Isolation of Lycotoxins-In the process of screening spider venoms for calcium channel antagonism, we found that venom of the wolf spider L. carolinensis (1:1000 dilution) disrupts 45 Ca 2ϩ ion flux in depolarized rat brain synaptosomes. As we demonstrate below, this effect results from dissipation of chemical and electrical gradients across membranes rather than from calcium channel antagonism. To further characterize this biological activity, we fractionated the venom by reversedphase liquid chromatography using the protocols described under "Experimental Procedures." Two fractions were identified and named lycotoxins I and II (Fig. 1A). Each lycotoxin was purified in subsequent steps to homogeneity (Fig. 1, B and C).
Calcium flux assays used to follow the purification of the lycotoxins are shown in Fig. 2. Depolarization of rat synaptosomes by elevating the external potassium concentration increases synaptosomal 45 Ca 2ϩ levels by almost 2-4-fold, due to opening of voltage-dependent calcium channels. Depolarization-induced 45 Ca 2ϩ uptake was completely inhibited following pre-exposure to lycotoxin I (10 M) or lycotoxin II (10 M) (Fig.  2). Both lycotoxins (10 M) also depleted basal 45 Ca 2ϩ levels in the absence of potassium depolarization (Fig. 2). Taken together, these results suggest that the lycotoxins interfere with the ability of synaptosomes to sequester Ca 2ϩ .
Amino Acid and Mass Analyses-Quantitative amino acid composition analysis of each lycotoxin gave the results shown in Table I; the presence of tryptophan was deduced from on-line UV absorption spectra of lycotoxins (data not shown). Amino acid composition analysis allowed us to determine the concentrations of lycotoxins in the crude venom to be ϳ5 mM. Preliminary mass analysis (matrix-assisted laser desorption time-of-    (Table II). The sequence of lycotoxin II was determined to be KIKWFKTMK-SIAKFIAKEQMKKHLGGE-OH (Table III). The deduced sequences are in good agreement with the results from amino acid composition analyses (Table I). The C-terminal amidation of lycotoxin I was assigned by high resolution electrospray mass spectrometry. The mass of the native peptide was [MH] ϩ ϭ 2843.4, which is virtually identical to the mass of 2843.5 predicted for the C-terminal amidated peptide, whereas that predicted for the free acid is 2844.5. Mass analysis of lycotoxin II yielded [MH] ϩ ϭ 3206.0, compared with the predicted values of 3205.8 for the free acid form and 3204.8 for the amidated form. From this comparison, we concluded that lycotoxin II occurs as the free acid.
To confirm the chemical identity of lycotoxin I and to provide sufficient quantities for further biological experiments, the peptide was prepared synthetically. Co-injection of native lycotoxin I (4 nmol) and synthetic lycotoxin I (3 nmol) gave a single peak on reversed-phase liquid chromatography (Fig. 1D); integration of this peak indicated the presence of both substances as the peak area was proportional to the total amount of the two materials tested. This confirmed that native lycotoxin I and synthetic lycotoxin I have identical RPLC elution profiles. Additionally, native and synthetic peptides were compared and found to be indistinguishable in their ability to deplete 45 Ca 2ϩ levels in rat synaptosomes (Fig. 3). Logistic curves fitted to the data showed that synthetic lycotoxin I had an EC 50 of 5.0 Ϯ 0.13 M as compared with native lycotoxin I with an EC 50 of 4.5 Ϯ 0.14 M. Taken together, these results show that synthetic lycotoxin I is chemically and biologically indistinguishable from the native peptide.
Amphipathic ␣-Helical Character of Lycotoxins-The primary structures of lycotoxins are characterized by lysine re -FIG. 4. A, sequence alignment of lycotoxins I and II with the ␣-helical amphipathic peptides magainin B and adenoregulin, all of which show a pattern of repeating lysine residues. B, ␣-helical wheel plot of lycotoxin I showing amphipathic character. In this conformation, periodic variation in the hydrophobicity value of the residues along the peptide backbone with a 3.6 residues/cycle period characterizes an ␣-helix (20). The amino acid sequence of lycotoxin I begins with isoleucine (I) and proceeds in a clockwise direction. Note that all lysine residues with the exception of one occur on the "hydrophilic side" (upper left) of the helix. C, ␣-helical wheel plot of lycotoxin II. In this plot, the hydrophilic side of the predicted ␣-helix lies at the lower right. Six positively charged lysine residues occur on the predicted hydrophilic side of the helix, and two occur on the hydrophobic side. peats occurring every fourth or fifth position. A GenBank™ search resulted in the identification of several peptides showing a similar lysine motif, including adenoregulin, a peptide from the skin secretion of the frog Phyllomedusa bicolor (17), and dermaseptins and magainins (14,18) (Fig. 4A). Magainins, dermaseptins, and adenoregulin are all antimicrobial poreforming peptides with characteristic amphipathic ␣-helical secondary structures (14,17,18). Interestingly, the predicted secondary structures of lycotoxins I and II show that the majority of their amino acids occur in an ␣-helix conformation according to  and Garnier-Osguthorpe-Robson (20) principles. When plotted as ␣-helical wheels, the majority of the hydrophobic and hydrophilic amino acid residues occur on opposite sides of the helix (Fig. 4, B and C). For lycotoxin I, for instance, six charged or hydrophilic residues occur on one side of the cylindrical surface, and eight hydrophobic residues are arranged on the opposite site. Except for the C-terminal lysine (lysine 24), all positively charged lysine residues are on the hydrophilic side. For lycotoxin II, six lysines and the C-terminal glutamate occur on the hydrophilic side, whereas eight hydrophobic residues and two lysines occur on the predicted hydrophobic side. This suggests that both lycotoxins I and II may be configured as amphipathic ␣-helices spanning large portions of the peptides. In fact, such amphipathic ␣-helical secondary structure has been predicted for many other peptides including PGLa, XPF, and bombinins from frog skin and cecropins from insects (21,22). The finding that lycotoxins are significantly conserved with antimicrobial peptides in both the primary and secondary structures led to experiments designed to test them for antimicrobial activity.
Antimicrobial Activity of Lycotoxins-The lycotoxins show potent antimicrobial activity against both prokaryotic and eukaryotic cells in plate growth inhibition assays. Against E. coli strain DH5 (Fig. 5, rows A-C), lycotoxin II was most active, showing inhibitory activity at concentrations as low as 40 M as compared with 60 -80 M for magainin B and 80 -150 M for lycotoxin I. The most sensitive to lycotoxin I was a Grampositive bacterial species, B. thuringiensis israelensis (Fig. 5, row E), for which the minimal inhibitory concentration of lycotoxin I was 5 M (lowest concentration tested); the minimal inhibitory concentration for magainin B and lycotoxin II was 60 M (Fig. 5, rows D and F). Against Gram-negative E. coli strain D31 (23), lycotoxins I and II were similarly active in the 10 -20 M range (Table IV). Finally, lycotoxins I and II inhibited growth of a yeast species, C. albicans, at 40 M, whereas twice that concentration was required for observable activity with magainin B (Fig. 5, rows G-I). A summary of the results from plate growth inhibition assays is shown in Table IV. In general, the lycotoxins were more active as antimicrobial agents than magainin B.
When tested in a liquid growth inhibition assay (Fig. 6), the antimicrobial actions of lycotoxin I, lycotoxin II, and magainin B appeared very abruptly at submicromolar concentrations (ϳ0.6 -0.7 M). When applied jointly, the effects of lycotoxins I and II were additive, suggesting that they have little or no synergistic action. The slope factors for these inhibition curves were in excess of 25; the steep relationship between concentration and lytic activity indicates a phase change in the membrane that most likely involves a high degree of cooperativity between toxin molecules and is characteristic of peptides for which aggregation confers antimicrobial activity (24).
Lysis of Erythrocytes-Pore-forming peptides such as the magainins lyse erythrocytes at high micromolar concentrations (13). Given the sequence similarities between the lycotoxins and magainins, we compared lycotoxin I and magainin B for the ability to lyse rabbit erythrocytes (Fig. 7). Neither lycotoxin I nor magainin B showed detectable hemolytic activity at concentrations below 30 M. However, both peptides showed significant hemolytic activity at concentrations above 100 M, with lycotoxin I being more active. For instance, at 200 M, 55% hemolysis was observed for lycotoxin I ((A lycotoxin /A H 2 O) ϫ 100%) and 35% for magainin B. These results provide further evidence for the pore-forming activity of lycotoxin I.
Lycotoxin Dissipates Ion Gradients-To further characterize the effects of lycotoxins on ion gradients, we performed several experiments using synaptosomes preloaded with 45 Ca 2ϩ (Fig.  8). Synaptosomes (ϳ180 g of membrane protein) accumulated a maximum of 65 pmol of 45 Ca 2ϩ within 10 min (Fig. 8, Treatments 1 and 3). Note that Treatment 2, which involved exposure to 45 Ca 2ϩ and cobalt simultaneously, blocked 45 Ca 2ϩ accumulation, whereas addition of cobalt after a 10-min delay (Treatment 3) permitted accumulation to the control level. Cobalt is known to antagonize both inward and outward 45 Ca 2ϩ movements due to inhibition of calcium channels and the Na ϩ -Ca 2ϩ exchanger (25).
Addition of synthetic lycotoxin I at time 0 (Fig. 8, Treatment  4) or after a 15-min preloading period (Treatment 5) reduced control 45 Ca 2ϩ accumulation by 80%. The result of Treatment 4 suggests that the toxin prevents 45 Ca 2ϩ sequestration, whereas the result of Treatment 5 indicates that the toxin promotes 45 Ca 2ϩ efflux from preloaded synaptosomes. If preloaded synaptosomes were treated sequentially with cobalt followed by lycotoxin I, the control response was reduced 50%. Taken together, these results suggest that lycotoxin I dissipates 45 Ca 2ϩ gradients in preloaded synaptosomes.
Lycotoxin I Dissipates Membrane Electrochemical Potential-Since insects are the natural prey targeted by wolf spiders, we tested the effects of lycotoxin I on insect musculature. Specifically, we examined the effects of the toxin on muscle membrane potentials. A pore-forming activity of lycotoxin would be manifested as a diminution of the resting membrane potential. Intracellular recordings were made from body wall muscles 6A and 7A of larval third instar house flies (M. domestica). The resting potential (E m ) and amplitude of the neurally evoked excitatory junctional potential (EJP) were recorded and plotted every 4 s (Fig. 9). Brief exposure to synthetic lycotoxin I (1 M) caused a slight but reversible depolarization of the E m and a corresponding decrease in the amplitude of the EJP. When 3 M lycotoxin I was applied, the E m depolarized from Ϫ60 to 0 mV. At the same time, the EJP was completely blocked. These results show that lycotoxin I dissipates the muscle cell resting potential in this preparation. DISCUSSION We have identified and characterized the lycotoxins, two broad-spectrum antimicrobial peptides that occur in venom of the wolf spider L. carolinensis. To our knowledge, this is the first report documenting the occurrence of antimicrobial peptides in spider venom. The lycotoxins occur at concentrations of 1-5 mM in whole L. carolinensis venom, which is in the range previously reported for agatoxins occurring in funnel web spider (Agelenopsis aperta) venom (12, 26 -28). Such high concentrations suggest that the lycotoxins play an important role in the biochemical strategy used by the spider to capture insect prey. We observed a complete loss of cell membrane potential and block of neuromuscular transmission in insect body wall muscles by lycotoxin I (3 M). Furthermore, the lycotoxins (2-20 M) caused efflux of calcium ions from rat brain synap-tosomes. These results indicate that the lycotoxins dissipate ion and voltage gradients across the membranes of excitable cells and likely contribute to paralysis of envenomated prey.
The utility of the lycotoxins in prey capture is complemented by their antimicrobial activity, which may serve as a defense against infectious organisms arising from prey ingestion. Precedent for this concept can be found in reports of magainins and defensins expressed by epithelial cells in gastric and intestinal mucosa (29 -31), where they may have multiple functions, including defense against microbial infection and regulation of symbiotic gut flora (32)(33)(34)(35). The mode of action of the lycotoxins therefore may serve two functions for the wolf spider: as a paralytic agent used for prey capture as well as a defense against infectious microbes arising from prey ingestion.
Lycotoxins I and II are cationic peptides containing 25 and 27 amino acids, respectively. They share 52% sequence similarity with each other. Theoretical predictions of their secondary structures suggest the formation of a cationic amphipathic ␣-helix in which lysine residues are clustered on the face of the cylindrical surface. This type of secondary structure is characteristic of magainins, dermaseptins, and adenoregulin. Indeed, lycotoxin I shows 60% sequence similarity to magainin B and 48% similarity to adenoregulin. Lycotoxin II shows 45% similarity to magainin B and 48% similarity to adenoregulin. The  antimicrobial actions of the lycotoxins are also consistent with this group of amphipathic ␣-helical peptides, which also includes PGLa, CPF, and XPF (36, 37).
Natural antimicrobial peptides with amphipathic ␣-helical structures (38) have been shown to disrupt biological membranes through pore formation (39 -45) or, in some cases, through destabilization of membrane phospholipid packing (46). These peptides form channels by self-aggregation of peptide monomers, whereby hydrophilic residues on one side of the helix face inward and hydrophobic residues on the opposite side of the helix interact with fatty acid side chains of the lipid bilayer.
At least four lines of evidence presented in this paper support the hypothesis that lycotoxins behave as pore-forming peptides. 1) They inhibit the growth of bacteria at a characteristic critical concentration, an indicator of self-aggregation of the peptide; 2) cause 45 Ca 2ϩ efflux from rat brain synaptosomes; 3) induce hemolysis of erythrocytes; and 4) dissipate the electrochemical potential across muscle membrane. Experiments in which lycotoxins showed antimicrobial activity were conducted in parallel with magainin B, which showed similar potency in some of the assays. Of particular interest are the data on lysis of E. coli strain D31, in which both lycotoxins I and II and magainin B showed extremely steep concentration response curves (Fig. 6). In contrast, these agents lysed red blood cells at much higher concentrations with parallel but far less steep concentration curves (Fig. 7). These data indicate that the lycotoxins and magainin B operate according to similar but probably distinct mechanisms in each system. Finally, the amino acid sequences of lycotoxin I and magainin B are quite similar, both being polar basic peptides with ␣-helical amphipathic character. Since magainin B belongs to a well established family of pore-forming peptides (41,43), we believe that the lycotoxins permeabilize cell membranes through similar mechanisms.
Although the lycotoxins share significant sequence similarity with magainin B, different specificities were observed for each peptide. Lycotoxin I was more effective against the Grampositive bacteria B. thuringiensis and yeast than magainin B. Magainin B (10 M) also showed no effect on synaptosomal Ca 2ϩ efflux (data not shown), whereas both lycotoxins caused 45 Ca 2ϩ efflux from synaptosomes.
The lycotoxins define a new functional subclass of spider venom toxins based on their amino acid sequences and antimicrobial activity. As with the magainins, we believe that the lycotoxins have the potential to stimulate development of new useful therapeutic agents (47,48). The rapid development of antibiotic-resistant bacteria points to the ongoing need for novel antimicrobial agents (38). Our findings suggest that spider venoms represent a potentially widespread source of novel antimicrobial agents that may have important medical implications. composition and protein sequencing analyses; and Tom Prentice for expertise in obtaining wolf spider venom. We are also grateful to Dr. Guoqiang Jiang for help in the antimicrobial inhibition assays and Dr. Janette Mills for valuable discussions during the preparation of this manuscript.