Plicatamide, an antimicrobial octapeptide from Styela plicata hemocytes.

Plicatamide (Phe-Phe-His-Leu-His-Phe-His-dc Delta DOPA), where dc Delta DOPA represents decarboxy-(E)-alpha,beta-dehydro-3,4-dihydroxyphenylalanine, is a potently antimicrobial octapeptide from the blood cells of the solitary tunicate, Styela plicata. Wild type and methicillin-resistant Staphylococcus aureus (MRSA) responded to plicatamide exposure with a massive potassium efflux that began within seconds. Soon thereafter, treated bacteria largely ceased consuming oxygen, and most became nonviable. Native plicatamide also formed cation-selective channels in model lipid bilayers composed of bacterial lipids. Methicillin-resistant S. aureus treated with plicatamide for 5 min contained prominent mesosomes as well as multiple, small dome-shaped surface protrusions that suggested the involvement of osmotic forces in its antimicrobial effects. To ascertain the contribution of the C-terminal dc Delta DOPA residue to antimicrobial activity, we synthesized several analogues of plicatamide that lacked it. One of these peptides, PL-101 (Phe-Phe-His-Leu-His-Phe-His-Tyr-amide), closely resembled native plicatamide in its antimicrobial activity and its ability to induce potassium efflux. Plicatamide was potently hemolytic for human red blood cells but did not lyse ovine erythrocytes. The small size, rapid action, and potent anti-staphylococcal activity of plicatamide and PL-101 make them intriguing subjects for future antimicrobial peptide design.

Phe-Phe-His-Leu-His-Phe-His-dc⌬DOPA (plicatamide) 1 is a modified octapeptide found in the hemocytes of Styela plicata (1). In the preceding sequence, dc⌬DOPA indicates decarboxy-(E)-␣,␤-dehydro-3,4-dihydroxyphenylalanine. Although the sequence of plicatamide did not resemble a conventional antimicrobial peptide, we examined its antimicrobial properties because hemocytes are key participants in innate antimicrobial defenses. Despite its small size, plicatamide proved to be a potent, rapidly acting, and broad spectrum antimicrobial. We also prepared the following four synthetic analogues that differed from plicatamide only in their C-terminal residue: tyrosine amide in PL-101; tyrosine acid in PL-102; DOPA (3,4dihydroxyphenylalanine) acid in PL-103; and DOPA-amide in PL-104. Of these octapeptides, PL-101 most closely simulated the antimicrobial properties of native plicatamide. This report will describe the effects of plicatamide on staphylococci.

Peptide Purification
Native plicatamide was purified from freshly harvested hemocytes (blood cells) of S. plicata as described recently (1). We determined their peptide content either by performing quantitative amino acid analysis or by doing analytical reverse phase-HPLC on a C18 column and then computing and comparing the area under the curve (AUC) at 215 nm with the AUC of an appropriate standard previously subjected to quantitative amino acid analysis.

Peptide Synthesis
The synthetic peptides used in our initial experiments were customsynthesized by Fmoc (N-(9-fluorenyl)methoxycarbonyl) chemistry at Research Genetics (Huntsville, AL) and purified to homogeneity by reverse phase-HPLC. Mushroom tyrosinase (6680 units/mg) was purchased from Sigma, and all other reagents were of analytical grade. Mushroom tyrosinase (Sigma) was used to prepare PL-103 and -104 (Table I) by converting the C-terminal tyrosine of PL-101 and -102 to DOPA (2,3). Briefly, the synthetic peptides (1 mg/ml, final concentration) were dissolved in 10 ml of 20 mM borate, 0.1 M phosphate/ascorbate buffer, pH 7.0, in a plastic reaction vessel. Before starting the reaction by adding 100 g/ml (final concentration) of mushroom tyrosinase (Sigma, 6680 units/mg), we removed a 50-l aliquot and acidified it with 2 l of 6 N HCl. Tyrosinase reactions were run at room temperature under a stream of humidified air. Aliquots of the reaction mixture were removed every 20 min and subjected to analytical reverse phase-HPLC to monitor the progress of the reaction. This chromatography was performed over 50 min on a Phenomenex Jupiter Series 4.6 ϫ 250-mm analytical C-18 column (10 M, 300-Å pore size), using a 0 -40% linear gradient of water with 0.1% trifluoroacetic acid to acetonitrile in 0.085% trifluoroacetic acid. After 60 min, the reaction was terminated by adding 200 l of 6 N HCl, and the mixture was desalted by loading it directly onto a Sep-Pac Vac 1-g (6 ml) cartridge (Waters Associates, Milford, MA). After washing the cartridge with 20 ml of water containing 0.1% trifluoroacetic acid, the peptides were eluted with 10 ml of 60% acetonitrile containing 0.085% trifluoroacetic acid. Subsequent purification was obtained by multiple runs on a 10 ϫ 250-mm C-18 reverse phase-HPLC column. The peptide sequences were checked by tandem mass spectrometry on a Finnigan LCQ Ion Trap Instrument. Subsequent batches of PL-101 were synthesized in our UCLA laboratory on an ABI 433A peptide * This work was supported in part by awards from the Stein-Oppenheimer Endowment Fund and the National Sea Grant Program of the United States Department of Commerce, National Oceanic Atmospheric Administration Grant R/MP-93 (through the California Sea Grant College Program and the California State Resources Agency). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ Preliminary studies were supported by a generous donation from the Kieckhefer Foundation.

Antimicrobial Assays
Radial Diffusion Assays-The assay has been described elsewhere (4). Our Gram-positive test organisms were Staphylococcus aureus 930918-3, MRSA ATCC 33591, a methicillin-resistant S. aureus strain, and Listeria monocytogenes, strain EGD. In some experiments we also tested Escherichia coli, ML-35p, and Pseudomonas aeruginosa, MR3007, a strain that was resistant to many conventional antibiotics. Native plicatamide was serially 3.16-fold diluted with 0.01% acetic acid that contained 0.1% human serum albumin to minimize its nonspecific adsorption to plastic tubes. Organisms were grown to mid-logarithmic phase at 37°C in trypticase soy broth. After they were washed with 10 mM phosphate buffer, pH 7.4, ϳ4 ϫ 10 6 bacterial colony-forming units (CFU) were incorporated into 10 ml of the underlay gel mixture. Unless otherwise stated, the underlay gels also contained 1% w/v agarose (Sigma A-6013), 10 mM sodium phosphate buffer, pH 7.4, and 0.3 mg/ml trypticase soy broth powder. Some underlay gels were supplemented with 100, 175, or 250 mM NaCl. A 6 ϫ 6 array of sample wells, each 3.2 mm in diameter and 1.2 mm deep, was punched in the underlay gel. These allowed 8-l aliquots of each dilution to be introduced. After the plates had incubated for 3 h at 37°C, a nutrient-rich overlay gel (60 mg/ml trypticase soy broth powder in 1% v/v agarose) was poured, and the incubation was continued overnight to allow surviving organisms to form microcolonies. The diameters of completely clear zones were measured to the nearest 0.1 mm and expressed in units (1 unit ϭ 0.1 mm), after first subtracting the well diameter. Because a linear relationship exists between the zone diameter and the log 10 of the peptide concentration, the X intercept of this line was determined by a least mean squares fit and was considered to represent the minimal effective concentration (MEC).
Colony Counting Assays-Stationary or mid-logarithmic phase bacteria were prepared as described above and incubated with antimicrobial peptides at 37°C in an agarose-free liquid medium containing 100 mM NaCl, 10 mM sodium phosphate, or Tris buffer, pH 7.4, and such other additions as are described in the text. Aliquots (20 l) were removed at intervals, diluted appropriately, and transferred to nutrient agar plates with a Spiral Plater (Spiral Biotech, Rockville, MD). Colonies were counted after overnight incubation at 37°C.
Broth Microdilution Assays-These assays used cation-adjusted, Mueller Hinton II Broth (BD Biosciences) and were performed according to the guidelines of the National Committee for Clinical Laboratory Standards (5), except that the 10ϫ stock plicatamide was prepared and serially diluted in acidified water (sterile 0.01% acetic acid) instead of in Mueller Hinton II Broth.

Potassium Release
Test organisms were incubated overnight in 50 ml of trypticase soy broth at 37°C, washed three times with buffer (100 mM NaCl, 10 mM Tris acetate, pH 7.4), and resuspended in this buffer at ϳ2.5 ϫ 10 8 CFU/ml, based on A 620 . Experiments were done at 37°C in stirred polypropylene tubes surrounded by a 50-ml water-jacketed reaction vessel (Kimble/Kontes, Vineland, NJ). The tube contained 6 ϫ 10 7 CFU of washed, stationary phase bacteria in 100 mM NaCl, 10 mM Tris acetate, pH 7.4, in a final volume of 250 l. An Orion SensorLink PCM-700 pH/ISE meter, fitted with a MI-442 potassium electrode (Microelectrodes, Bedford, NH) and an SDR-2 reference electrode (World Precision Instruments, Sarasota, FL), was used as described previously (6).

Oxygen Consumption
Oxygen consumption by washed, stationary phase bacteria was measured with a Clark-type oxygen electrode (Hansatech Ltd., Norfolk, UK). Briefly, an overnight culture of MRSA was twice washed with PBS and adjusted to 3 ϫ 10 8 CFU/ml in this medium. After adding bacteria (1 ml) to the continuously stirred chamber, we added 10 l of full strength trypticase soy broth, and we measured the basal rate of O 2 consumption at room temperature for 5-10 min before adding peptide (10 g/ml final concentration).

Cytotoxic and Hemolytic Activity
The cytotoxicity of plicatamide for ME-180 (ATCC HTB-33) human cervical epithelial target cells was assessed with an MTT-tetrazolium reduction assay (Roche Molecular Biochemicals). Target cells were grown to confluency in RPMI 1640 medium with 10% fetal bovine serum and 50 g/ml gentamicin and harvested with trypsin/EDTA. After washing them with this medium, their concentration and viability (trypan blue exclusion) was determined, and they were suspended at 5 ϫ 10 4 cells/ml. Cell aliquots (100 l) were dispensed into 96-well tissue plates (Corning Glass) and incubated for 5 h at 37°C in room air with 5% CO 2 before the peptides were added. After 20 additional hours of incubation, we added 10 l of MTT solution, followed 4 h later by 100 l of extraction buffer. After overnight extraction of the reduced MTTtetrazolium, absorbance was measured at 600 and 650 nm, on a Spectramax 250 spectrophotometer (Molecular Devices, Sunnyvale, CA).
Hemolytic activity was tested by incubating various concentrations of peptide with a suspension (2.8% v/v) of washed human or sheep red cells in Dulbecco's phosphate-buffered saline. After 30 min at 37°C, the tubes were centrifuged, and the absorbance (A) of the supernatants was measured. The percentage of hemolysis was calculated by Equation 1, where A exper and A control signify the absorbance values of supernatants from treated and untreated red cells, and A total is the supernatant of red cells treated with 0.1% Triton X-100.

Electron Microscopy
For transmission electron microscopy, 5 ϫ 10 8 bacterial CFU/ml were exposed at room temperature to 42.5 g/ml native plicatamide in PBS (100 mM NaCl and 10 mM sodium phosphate, pH 7.4) containing 1% v/v trypticase soy broth. At intervals, 1-ml aliquots were removed, centrifuged briefly at 2000 ϫ g, and immediately resuspended in 1 ml of freshly made 2% glutaraldehyde in PBS. After 30 min on ice, the fixed organisms were washed in PBS.
For scanning EM, 10% of the above bacteria were adhered for 30 min to mixed cellulose ester membrane filters with 0.025-m pores (Millipore, Bedford, MA). The filters were washed twice with 10 mM sodium phosphate, pH 7.4, and dehydrated through a graded ethanol series into hexamethyldisilane. After carbon coating, the samples were viewed on a Cambridge Stereoscan Electron Microscope.
The remaining bacteria were washed in PBS, post-fixed for 45 min at room temperature in 1% osmium tetroxide, dehydrated through ethanol to propylene oxide, and embedded in Epon 812. After staining with uranyl acetate at 60°C for 15 min, and then by lead citrate, the sections were viewed on a JEOL CX II microscope.

Planar Lipid Bilayers
Solvent-containing phospholipid bilayer membranes were formed by placing a small bubble of 15 mg/ml lipid solution in n-heptane onto the end of Teflon tubing with 0.25-mm inner diameter. The design of the chamber allowed 50 l of solution to be rapidly introduced immediately adjacent to the membrane (7). E. coli total lipid extract were purchased from Avanti Polar Lipids (Alabaster, AL) and stored at Ϫ20°C. Agar salt bridges connected the electrodes to the solutions, and voltage clamp conditions were employed in all experiments. The cis-side (i.e. the side to which peptide was added) was taken as ground. All stated voltages refer to the voltage of the trans-side. Current was recorded with an Axopatch-1C amplifier with a CV-3B head stage and stored on videotape for later playback and analysis. Membrane capacitance and resistance were monitored to ensure the formation of reproducible membranes. The peptide stock solution (2 mg/ml) was stored at 4°C, and the working solutions were prepared immediately before use. The bath solution contained 100 mM KCl and 10 mM Tris-HCl buffer, pH 7.4, or 10 mM MES-Tris buffer, pH 5.5 and pH 6.5, or 10 mM Tris citrate buffer pH 7.4.

RESULTS
Antimicrobial Activity of Plicatamide- Fig. 1a summarizes a series of radial diffusion assays done in underlay gels containing 100 mM NaCl at pH 7.4 and pH 5.5. Both native plicatamide and PL-101 (Phe-Phe-His-Leu-His-Phe-His-Tyr-amide) were more effective microbicides at neutral pH. Although native plicatamide and PL-101 had similar potency against the two Gram-positive organisms, S. aureus and L. monocytogenes, the native plicatamide was 2-3-fold more potent against the Gramnegative test strains, E. coli and P. aeruginosa. Fig. 1b shows that PL-101 was considerably more active than either PL-102 (Phe-Phe-His-Leu-His-Phe-His-Tyr-acid) or the two DOPAcontaining peptides PL-103 and PL-104. The sequences of these peptides are shown in Table I.
Composition of Native Plicatamide Preparations-In several of our preparations of native plicatamide, FTIR analyses re-vealed additional bands, characteristic of lipids and/or phospholipid, in addition to the expected absorption bands for a peptide (Fig. 2a). A mixture of synthetic PL-101 and palmitoyloleoylphosphatidylglycerol provided a similar FTIR spectrum (Fig. 2b), whereas PL-101 gave a typical peptide spectrum (Fig. 2c). Because the antimicrobial data shown in Fig. 1a were obtained with a preparation of plicatamide that contained only the expected peptide bands, we consider it unlikely that any co-purified lipids were responsible for the antimicrobial properties of our other preparations. Fig. 3 shows the results of experiments comparing the effects of pH and salinity on the antimicrobial activity of plicatamide and PL-101. These native plicatamide preparations did contain associated (phospho)lipids by FTIR. Again, native plicatamide and synthetic PL-101 were substantially more effective at pH 7.4, than at pH 5.5 despite their greater cationicity at the lower pH. The MEC of plicatamide in 100 mM NaCl at pH 7.4 ranged from 1.0 to 2.5 g/ml for E. coli, S. aureus, and L. monocytogenes. These results are quite similar to those obtained with the phospholipid-free preparation of plicatamide (Fig. 1a). We obtained similar MEC values when the underlay gels contained 250 g/ml (data not shown).
Colony counting experiments revealed that native plicatamide killed MRSA and S. aureus very rapidly (Fig. 4). The peptide was equally effective in medium with or without nutrients, and we found little difference in the susceptibility of mid-logarithmic and stationary phase staphylococci to plicatamide (data not shown). Furthermore, staphylocidal activity was not impaired by including 10 g/ml catalase in the medium, nor did inclusion of 1 mM Ca 2ϩ or 1 mM Mg 2ϩ impair it (data not shown).
Effect on Bacterial Membrane Integrity-We assessed the membrane integrity of plicatamide-treated staphylococci by measuring their loss of cytoplasmic potassium (Fig. 5). To ensure adequate amounts (100 -200 nmol) of total K ϩ , bacterial concentrations of ϳ7.5 ϫ 10 7 CFU/ml were used. We also measured viability by removing aliquots at intervals for colony counting. The virtually immediate and substantial efflux of K ϩ from plicatamide-treated MRSA is consistent with an antimicrobial mechanism that targets their cell membrane. Native plicatamide induced a similar efflux of K ϩ from S. aureus, and synthetic PL-101 induced K ϩ efflux from both S. aureus and MRSA (data not shown).
Model Membrane Bilayers-We also examined the effects of plicatamide on planar bilayer membranes prepared from E. coli lipids dissolved in n-heptane. The untreated membranes were stable between Ϯ100 mV, and displayed low (Ͻ10 picosiemens) conductance. At pH 7.4, plicatamide concentrations below 5  Table I.  2. FTIR spectroscopy. a shows the spectrum of a preparation of purified native plicatamide. Its amide I and amide II bands are labeled, and the position of the bands characteristic of lipid acyl chains and phosphates are shown. b shows the spectrum of a 1:1 (by weight) mixture of palmitoyloleoylphosphatidylglycerol (POPG) and PL-101. c shows the spectrum of synthetic PL-101. Samples were twice freezedried from hexafluoroisopropanol, 10 mM HCl (7:1, v/v) prior to measurement to remove any (fluoro)acetate counterions that could interfere with the amide I spectral region (48). These (fluoro)acetate-free samples were solvated in hexafluoroisopropanol, spread on the germanium ATR surface, and dried under nitrogen before the spectrum was measured with a Bruker Vector 22 TM FTIR spectrometer. Spectra were averaged from 124 scans at a gain of 4 and a spectral resolution of 2 cm Ϫ1 . g/ml caused short spikes of increased conductance, whereas concentrations between 5 and 10 g/ml sometimes induced substantial conductivity (Fig. 6a), with an essentially linear integral current-voltage response (Fig. 6b). Plicatamide concentrations above 10 g/ml typically increased conductance very quickly before destroying the membrane. When these various concentrations of plicatamide were added at pH 5.5, no increased conductivity resulted (data not shown).
The plicatamide-modified membranes were cation-selective, manifesting a current reversal potential of ϩ15.5 Ϯ 3.5 mV for a 10-fold KCl gradient. They showed relatively little selectivity for potassium versus sodium (2.1 Ϯ 1 mV for bi-ionic system at 100 mM, and 10.7 Ϯ 3.2 mV for 100 mM KCl, 1 M NaCl system). Although adding 1-10 mM CaCl 2 had no effect on plicatamideinduced currents, adding 1 mM ZnCl 2 blocked current flow by up to 75% (data not shown). Because we have found that plicatamide binds zinc, 2 we attribute the inhibitory effect of this cation to its interaction with plicatamide, rather than to nonspecific stabilization of the membrane. It is noteworthy that ZnCl 2 blocks pore formation by two histidine-rich polypeptides: aerolysin from Aeromonas hydrophila (8,9) and a histidine-rich analogue of staphylococcal ␣-hemolysin (10).
Oxygen Consumption-Exposing MRSA to plicatamide or PL-101 quickly decreased their consumption of oxygen by 87.1%, from a basal rate of 5.1 to 0.66 nmols/min/10 8 CFU, within 60 s after 10 g/ml native plicatamide was added (Fig.  7). Synthetic PL-101 was almost as effective, reducing O 2 consumption by 81.8% to 0.93 nmols/min/10 8 CFU. Untreated control organisms continued to consume O 2 at the basal rate until the chamber became anaerobic .
Microbroth Dilution Assays-We also performed conventional, National Committee for Clinical Laboratory Standards (NCCLS) microbroth dilution assays to determine the susceptibility of E. coli ML-35p, P. aeruginosa MR3007, S. aureus 930918-3, and L. monocytogenes EGD to native plicatamide and PL-101. In contrast to its prominent antimicrobial effects in our radial diffusion and colony count experiments, the MIC of plicatamide exceeded 100 g/ml for each of the aforementioned organisms. Because microbroth dilution assays are widely considered to represent "gold standards" in testing antimicrobials, we decided to investigate the cause of this apparent discrepancy. Initially, we suspected that the Mueller-Hinton broth used in National Committee for Clinical Laboratory Standards-type assays might not support plicatamide mediated staphylocidal activity. However, when we exposed mid-logarithmic or stationary phase S. aureus to 2 or 5 g/ml of plicatamide in Mueller-Hinton broth, the colony counts fell by Ͼ2-3 logs after 30 and 120 min of incubation (data not shown). A few additional experiments revealed that the few organisms that survived exposure to plicatamide could repopulate the culture, thereby masking the antimicrobial properties of plicatamide, at least for microbroth dilution assays. This effect is illustrated in Fig. 8.
Effects on Bacterial Ultrastructure-MRSA treated with plicatamide showed many alterations. After only 5 min, striking changes were observed by scanning electron microscopy, wherein multiple small dome-shaped bulges, often arranged in linear and clustered arrays (Fig. 9), deformed their surfaces. These abnormalities became more marked as the duration of exposure to plicatamide increased (Fig. 10). In many bacteria, large amounts of cytoplasm extruded beyond the confines of the cell wall. Transmission electron microscopy of plicatamide-treated bacteria revealed fixed prominent mesosomes, even in cells fixed as early as 5 min after exposure to plicatamide (Fig. 11). Many plicatamide-treated MRSA contained electron dense material between their plasma membrane and cell wall, representing partially contained "eruptions" of cytoplasm akin to the more flamboyant manifestations evident in Fig. 10.
Hemolytic and Cytotoxic Properties-Plicatamide-lysed human erythrocytes, acting with almost the same potency as melittin on a weight/volume basis (Fig. 12). However, in marked contrast to melittin, plicatamide was not hemolytic for sheep red blood cells, even when applied at 80 g/ml. Moreover, although melittin induced hemolysis over a broad pH range, the hemolytic properties of plicatamide were markedly diminished as acidity increased. Furthermore, whereas melittin was exceptionally cytotoxic for human cervical ME-180 epithelial cells, plicatamide was relatively noncytotoxic for these cells under the same conditions (data not shown). DISCUSSION Plicatamide (Fig. 13) is an interesting peptide for many reasons, not the least of which is that it violates conventional notions about antimicrobial peptides. Typically, one expects such peptides to be cationic and amphipathic molecules with 16 -40 residues (11)(12)(13)(14)(15). A few smaller antimicrobial peptides with 11-13 residues have been described. These include the bactenecin dodecapeptides of bovine or ovine neutrophils (16,17), bovine indolicidin (18,19), and tigerinins, antimicrobial peptides isolated from the skin secretions of a frog, Rana tigerina (20). Plicatamide contains eight residues, and it is only modestly cationic at pH 7.4, and when it was rendered more cationic (at pH 5.5) its activity decreased. To our knowledge, only two smaller antimicrobial peptides have been found in animals: 5-S-GAD, and halocyamine A. N-␤-alanyl-5-S-glutathionyl-3,4-dihydroxyphenylalanine (5-S-GAD) is a pentapeptide that was purified from the hemolymph of injured or infected "fleshflies" (Sarcophaga peregrina) (21). Because the antimicrobial activities of 5-S-GAD were completely inhibited by catalase, it was suggested that H 2 O 2 participates in its antimicrobial mechanism, as well as its induction of apoptosis in HL60 cells (22). Because we found that catalase did not inhibit the bactericidal effects of plicatamide, its antimicrobial mechanism evidently differs from that of 5-S-GAD. Other properties of the 5-S-GAD molecule include an ability to inhibit tyrosine phosphorylation of certain kinases, PTK p60(v-src) and PTK p210(BCR-ABL) (23,24).
Halocyamine A is a tetrapeptide (histidyl-3,4-dihydroxyphenylalanine-glycyl-bromodidehydrotryptamine) that, like plicatamide, is also found in the hemocytes of a tunicate, in this case Halocynthia roretzi. Halocyamine A was reported to inhibit the growth of yeast and of the marine bacteria Achromobacter aquamarinus and Pseudomonas perfectomarinus (25). Neither its antimicrobial mechanism nor the effects of catalase on its activity have been described.
It is remarkable that three of the smallest known antimicrobial peptides (5-S-GAD, halocyamine A, and plicatamide) should all contain a DOPA moiety. Although this could be a coincidence, it may also be an indication that this residue plays an important functional role. Although PL-103 and PL-104, the DOPA-containing synthetic analogues of plicatamide examined here, were unimpressive microbicides, we have yet to prepare an exact synthetic replica of plicatamide.
Another possible function of DOPA and dc⌬DOPA might be to impart adhesive properties that help retain the peptide at sites of injury or infection. Byssal threads and plaques, the major adhesive structures of marine mussels (Mytilus spp.), invariably contain DOPA (26). If a Lewis base and an oxidase such as polyphenol oxidase are both present, DOPA can be converted to a DOPA quinone, whose spontaneous tautomerization forms ␣,␤-dehydro-DOPA (27,28).
Tunicates are protochordates-invertebrates that belong to the phylum chordata. Although the functional biochemistry of tunicate hemocytes has received relatively little attention, much is known about the microbicidal mechanisms of mammalian white blood cells, especially polymorphonucleated granulocytes (PMN). Mammalian PMN employ two principal strategies to kill microorganisms. One strategy involves using an array of antimicrobial peptides and proteins (29), and the other depends on the production of oxidants by postphagocytic metabolism (30). The principal oxidants of human PMN are produced by an NADPH oxidase complex (31,32) and include H 2 O 2 , OH (hydroxyl radical), and "downstream" products such as chloramines and hypochlorous acid formed by interactions between H 2 O 2 and myeloperoxidase (33). In the PMN of rodents and some other mammals, copious amounts of nitric oxide are formed by an inducible nitricoxide synthase (34).
Although it is not known if tunicate hemocytes have NADPH oxidase or inducible nitric-oxide synthase activity, phenol oxidase is released from the hemocytes of H. roretzi after foreign cells (e.g. yeast) are encountered. H. roretzi phenol oxidase, a metalloenzyme that requires copper ions for full activity (35), was reported to be antibacterial in the presence of DOPA and H. roretzi hemolymph. Because certain hemocytes ("morula cells") of the colonial ascidian Botryllus schlosseri also contain More advanced surface deformities were noted in MRSA exposed to native plicatamide (15 g/ml) for 15-60 min. In addition, many cells showed transmural extrusions of their cytoplasm.
FIG. 11. Transmission electron microscopy. Many plicatamidetreated MRSA contained mesosomes (white arrows), and such structures were evident as early as 5 min post-exposure to plicatamide. After longer exposures to plicatamide (15 g/ml), some MRSA contained electron dense material between their plasma membrane and thick wall.
phenoloxidase (36), this enzyme may be widely distributed in tunicates. The possibility that the DOPA moiety of plicatamide endows the molecule with an ability to participate in oxidative microbicidal reactions deserves consideration, especially because its histidines and modified DOPA residue endow plicatamide with the ability to bind transition metals. 2 Whereas direct measurements of potassium efflux have seldom been applied to antimicrobial peptides, many investigators have used membrane potential sensitive carbocyanine dyes to follow their effects on bacterial membrane potential. For example, in a recent study of S. aureus and S. epidermidis, Hancock and co-workers (37) compared the kinetics of killing (by colony counts) with that of membrane depolarization. At early time points, when membrane depolarization was incomplete, 90% or more of the bacteria had been killed. These results are relevant to our findings with plicatamide, because membrane potential and intracellular potassium concentrations are related by the Nernst equation: E eq ϭ (RT/F) ln((K o )/(K i )). In this equation, E eq represents the membrane potential at equilibrium; (K o )/ (K i ) is the ratio of potassium concentrations outside and inside the cell; ln represents natural logarithm; T is the absolute temperature; R is the universal gas constant; and F (the Faraday) is a physical constant. In our studies, the fate of plicatamide-treated bacteria appeared to be determined as soon as the process responsible for their potassium loss began, which should correlate with the onset of the depolarization process.
The ability of plicatamide to induce a massive potassium efflux from staphylococci suggests that it acts on their plasma membrane. The rapidity of its lethal and leakage effects prove that the thick staphylococcal cell wall peptidoglycan is not a barrier to the diffusion of plicatamide. The decreased activity of plicatamide under acidic conditions (pH 5.5), when its net positive charge would be highest, suggests that electrostatic interactions (e.g. with anionic phospholipids of the bacterial membrane) are unlikely to play a major role in its activity. The data reported here show that plicatamide retains its antimicrobial activity in our "high salt" conditions (250 mM). More significantly, at least from the perspective of a tunicate, plicatamide retained full activity in media containing 450 mM NaCl (i.e. the NaCl concentration of seawater) and in seawater itself (data not shown).
The striking morphological changes in plicatamide-treated MRSA somewhat resemble the membrane "blebbing" often shown by Gram-negative bacteria after exposure to antimicrobial peptides (38,39). However, because Gram-positive bacteria lack an outer membrane and encase their cytoplasmic membrane within a thick cell wall, these changes have a different genesis. We interpret the bleb-like protrusions of plicatamidetreated MRSA as osmotically driven herniations of the plasma membrane through small clefts ("tesserae") in the peptidoglycan fabric of the cell wall. We have seen similar changes with other antimicrobial peptides, especially protegrins (40), and will describe them in detail elsewhere.
Whereas ovine erythrocytes were resistant to lysis by plicatamide (Fig. 12), human erythrocytes were susceptible. Although the reasons for this difference are unclear, very similar findings have been noted with respect to lysis by other antimicrobial peptides (SMAP-29 and protegrin PG-1) (13,41), bile salts (42), or hypotonic conditions (43). Several factors may contribute to the differential susceptibility of human and sheep erythrocytes to lysis. Among these are their marked difference in size (sheep red cells are considerably smaller) and phospholipid composition (sheep erythrocytes lack phosphatidylcholine) (44,45). In addition, some strains of sheep have red blood cells whose intracellular ionic composition differs greatly from that of human erythrocytes (46).
From the standpoint of humans, the rapid and potent effects of plicatamide and PL-101 on staphylococci are also of interest. Infections caused by glycopeptide-resistant staphylococci and enterococci are becoming increasingly common, and additional agents that are effective against VanA strains of enterococci and "GISA"-type S. aureus are urgently needed (47). Although plicatamide was not especially cytotoxic and had little hemolytic activity for sheep erythrocytes, both plicatamide and PL-101 were extremely hemolytic for human red blood cells. We are currently trying to design active oligopeptide analogues of plicatamide with an improved cytotoxicity/hemolysis profile. If successful, these efforts could lead to practical applications.