NK-Lysin and Its Shortened Analog NK-2 Exhibit Potent Activities against Trypanosoma cruzi

Abstract

Antimicrobial peptides are widespread in nature and have been evolutionarily conserved as essential tools for combating a variety of pathogens. Among the plethora of natural peptides and synthetic analogs thereof studied in recent years for their antimicrobial activities, only a very few are known to be effective against protozoan parasites. In the present study we investigated the activity of NK-lysin, a broad-spectrum effector polypeptide of mammalian cytotoxic lymphocytes, against trypomastigotes of the human pathogen Trypanosoma cruzi in vitro. Moreover, the activity of a synthetic peptide named NK-2 that corresponds to the cationic core region of NK-lysin was tested in parallel against this parasite. T. cruzi was found to be highly susceptible to both peptides, as evidenced by inhibition of the mobility of trypomastigotes. The peptides rapidly permeabilized the plasma membrane of the parasite since micromolar concentrations resulted in the release of cytosolic enzymes within minutes. NK-lysin and NK-2 were even found to kill trypanosomes residing inside the human glioblastoma cell line 86HG39, but only NK-2 left the host cells apparently unharmed.

The protozoan parasite Trypanosoma cruzi infects a variety of mammalian species and is the etiologic agent of Chagas' disease of humans. Although it is a major health problem in Latin America (18 million people are chronically infected and 50,000 people die annually from Chagas' disease [27]), no convincing therapy is available. The presently used drugs nifurtimox and benznidazole are highly toxic and can prevent clinical manifestations only when patients are treated during the acute phase of the disease (21, 34). This has prompted several investigators to study the role of innate and acquired immunity in T. cruzi infections. In infected hosts, trypomastigotes circulate in the blood and invade a variety of cells, in which they multiply intracellularly as amastigotes (33). CD8⁺ T cells as well as CD4⁺ T cells are capable of restricting the growth of T. cruzi in a mouse model of infection (17, 22, 30), as it has been demonstrated for other intracellular pathogens. The protective capacity of CD4⁺ T cells relies on gamma interferon (IFN-γ)-mediated activation of macrophages and on providing help for antibody-producing B cells, whereas CD8⁺ T cells can directly lyse infected host cells. CD8⁺ T cells were shown to play a pivotal role in clearance of the pathogen. Induction of specific CD8⁺ T cells or adoptive transfer of these cells can induce protective immunity (17, 22). The antiparasitic effects of CD8⁺ T cells can be mediated by two different effector mechanisms: a perforin-independent mechanism mediated by IFN-γ and a perforin-dependent mechanism. Whereas IFN-γ stimulates macrophages to produce the trypanocidal molecule NO, the perforin-dependent mechanism is less well defined: on the one hand, perforin-knockout mice are highly susceptible to T. cruzi infection (24); on the other hand, the parasite itself is highly resistant to the lytic action of perforin (7).

This indicates that molecules from cytotoxic T lymphocytes other than perforin might be able to kill T. cruzi in synergy with the pore-forming protein. The most likely candidates are recently described membranolytic polypeptides found in cytoplasmic granules of mammalian cytotoxic T cells and natural killer (NK) cells (2, 26). The human form is termed granulysin, and the porcine homolog is named NK-lysin. Both are members of the family of saposin-like proteins, a group of proteins which share distinct structural features and phospholipid-interacting activity but fulfill diverse functions (16, 23, 28). It has been shown that granulysin and NK-lysin are antimicrobial and lyse a variety of pathogens (2, 32). Importantly, for granulysin it has been reported that it exhibits antimicrobial activity against intracellular pathogens in synergy with perforin, with which it is stored in cytoplasmic granules (32). A model has been proposed in which perforin and granulysin are transferred into the contact zone between the infected target cell and the effector cell (32). In this model, perforin allows granulysin to access its intracellular target. Released pathogens may be killed by the antimicrobial peptide to prevent infection of neighboring cells.

The finding that cytotoxic lymphocytes are equipped with antimicrobial polypeptides to participate in the host defense against intracellular pathogens is building a bridge between innate and acquired immunity (14). However, at high concentrations granulysin and NK-lysin also exhibit lytic activities against mammalian cells (2, 26), which limits their usefulness as therapeutic agents. Several attempts have been made to separate a structural element that is responsible for the antimicrobial activities of these relatively large peptides from others that are harmful for mammalian cells (3, 4, 10, 35). NK-2, a shortened synthetic peptide comprising residues 39 to 65 of NK-lysin, displays lytic activity against the fungal pathogen Candida albicans and a variety of gram-positive and gram-negative bacteria but exhibits virtually no hemolytic or cytotoxic activity against human cells (3).

This study was aimed at determining whether the antimicrobial spectrum of NK-lysin and its potent synthetic segment (residues 39 to 45), NK-2, may be extended toward the protozoan parasite T. cruzi. We compared the trypanocidal activities of NK-lysin and NK-2 by monitoring their effects on the viability and membrane integrity of the parasites. Moreover, we analyzed whether these peptides are capable of killing trypanosomes residing within host cells without affecting the viability of the latter.

MATERIALS AND METHODS

Peptides.

NK-lysin was expressed recombinantly in bacteria. The coding sequence of mature NK-lysin (GenBank accession no. X85431) was amplified by reverse transcription-PCR from RNA of porcine leukocytes and cloned into the vector PET28a (Novagen) by using NdeI and HindIII restriction sites which had been added to the 5′ ends of the sense and antisense oligonucleotides, respectively, used for reverse transcription-PCR. The expression was performed in Escherichia coli BL21(DE3) under isopropyl β-d-galactopyranoside induction (1 mM) for 2 h. The resulting inclusion bodies were harvested and solubilized, and the recombinantly expressed protein was purified by affinity chromatography on an Ni²⁺-nitrilotriacetic acid agarose column under denaturing conditions, according to the protocol of the manufacturer (Qiagen). Refolding was achieved by dialysis against decreasing concentrations of guanidine hydrochloride, addition of reduced and oxidized glutathione (10 and 1 mM, respectively), and subsequent dialysis against 20 mM acetic acid. After dialysis against phosphate-buffered saline (PBS) the six-His tag was cleaved by thrombin (Amersham Biosciences), and the released protein was purified by cation-exchange chromatography on a Resource S matrix (Amersham Biosciences) in 50 mM sodium acetate (pH 4.5) by using an NaCl gradient from 0 to 1 M. The eluted, pure protein was dialyzed against PBS.

NK-2 has been constructed according to the cationic core region of NK-lysin. It comprises residues 39 to 65 of the parent molecule (3) and was synthesized on commission by Affinity Research Products (Exeter, United Kingdom) by solid-phase techniques by a 9-fluorenylmethoxy carbonyl-based protecting group strategy. The sequence is KILRGVCKKIMRTFLRRISKDILTGKK with an amidated C terminus. The peptide was obtained in a high-purity grade (>95%). High-performance liquid chromatography-grade lyophilized melittin was purchased as a synthetic peptide from Sigma. NK-2 and melittin were dissolved in 0.01% trifluoroacetic acid. The concentrations of the NK-lysin and NK-2 peptides were determined by measuring the absorbance at 214 nm by using extinction coefficients derived from the respective sequence information (8).

Culture of parasites and host cells.

The macrophagotropic Tehuantepec strain of T. cruzi (29) was used in all experiments. Trypomastigotes were maintained in tissue culture by infection of monolayers of the human glioblastoma cell line 86HG39, grown in RPMI supplemented with 5% fetal calf serum (FCS) (6). Parasites released from their dying host cells 5 to 10 days later were collected and centrifuged at 100 × g for 5 min at 4°C to remove host cell debris. The parasites were sedimented by centrifugation at 500 × g for 5 min at 4°C and washed in RPMI supplemented with 5% FCS. For some experiments, lacZ-transfected trypomastigotes which expressed β-galactosidase in their cytosols were used (9).

Effects of peptides on T. cruzi and their host cells.

To determine the antitrypanosomal activities as well as the cytolytic activities of the peptides against the host cells, various concentrations of the respective peptides were incubated with 10⁵ T. cruzi or 10⁵ 86HG39 cells in 200 μl of RPMI supplemented with 5% FCS at 37°C for 30 min. The effect on T. cruzi trypomastigotes was directly visualized by light microscopy, and the concentration at which more than 50% of the parasites lost their motility was determined. The cytolytic effect on 86HG39 cells was evidenced by propidium iodide (PI) uptake by flow cytometry, and the concentration at which more than 50% of the cells were PI positive was determined.

Measurement of β-galactosidase release.

As a more sensitive assay, the release of a cytosolic β-galactosidase was measured. For this purpose, lacZ-transfected trypomastigotes (10⁶) in 50 μl of RPMI were incubated with the respective peptides at a concentration of 5 μM. At the indicated time points, parasites were spun down and the β-galactosidase activities of the supernatants were monitored as described previously (9). Briefly, the supernatants were incubated with 100 μM chlorophenol red-β-d-galactopyranoside (CPRG) in PBS at 37°C for 5 h, and the absorbance was measured at 560 nm by using a 96-well microtiter plate enzyme-linked immunosorbent assay reader. As a positive control trypomastigotes were lysed by addition of 1% Nonidet P-40 (NP-40).

Assay for trypanocidal activity.

To determine the effects of the different antimicrobial peptides on the infectivity of T. cruzi, 10⁴ parasites in 50 μl of RPMI supplemented with 5% FCS were incubated at 37°C for 3 h with different concentrations of the respective peptides. The complete mixtures were transferred to monolayers of 86HG39 cells in a flat-bottom 96-well microtiter plate and were incubated for various time periods. Afterwards, the cells were lysed by addition of 1% NP-40, and the β-galactosidase activity was measured as described above. By using calibrated amounts of lacZ-transfected parasites, the CPRG-specific absorption was found to be linear in relation to the numbers of parasites in culture.

Effects on intracellular T. cruzi.

To examine whether NK-2 and NK-lysin kill intracellular T. cruzi, monolayers of 86HG39 cells were infected with trypomastigotes as described above, and adherent pathogens were removed by several washes with medium. Infected monolayers were incubated for either 3 or 24 h with different concentrations of the respective peptides in RPMI supplemented with 5% FCS. After the indicated incubation period, the medium was removed and the remaining cells were washed and cultivated with fresh medium until a few free trypomastigotes were detectable in control cultures. Afterwards, the cells were lysed by addition of 1% NP-40 to the cultures. The β-galactosidase released by this treatment was measured to estimate the amount of intracellular T. cruzi in the culture.

Light microscopy.

T. cruzi trypomastigotes (10⁵) were incubated at 25°C for 30 min with the buffer control or with the respective peptides at a concentration of 5 μM in 200 μl of RPMI. The parasites were subsequently placed on glass slides, air dried, and stained with Giemsa.

Electron microscopy.

Trypomastigotes (10⁵) were incubated at 25°C for 30 min with the buffer or with the respective peptides at a concentration of 5 μM in 200 μl of RPMI. After 1 h the parasites were washed with PBS and fixed with 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer. The specimens were mounted on grids, sputtered with gold, and analyzed by scanning electron microscopy with a Philips SEM 500 scanning electron microscope.

Immunofluorescence.

86HG39 cells were grown on glass coverslips in 24-well plates. Confluent monolayers were infected overnight with T. cruzi at a ratio of one trypomastigote per 86HG39 cell. Adherent parasites were removed by washing, and the infected cells were subsequently incubated with the respective peptide in medium for 24 h. Afterwards, the cells were washed and cultured in fresh medium. The cells were cultured for an additional 72 h and were fixed with acetone. The samples were dried and stained with 4′,6′-diamidino-2-phenylindole (DAPI) for the detection of nuclei of both host cells and parasites, with fluorescein isothiocyanate (FITC)-labeled phalloidin for the detection of actin, and with an antiserum raised in rabbits against a soluble extract of T. cruzi for the detection of parasites.

Statistical analysis.

The results are presented as the means plus standard deviations. The number of individual experiments is indicated in each figure legend. Statistical analysis was generally performed by the unpaired Student's t test with the Prism software (Graph Pad Software, San Diego, Calif.). The level of significance was set at a P value of <0.05.

RESULTS

To compare the effects of NK-lysin and NK-2 (Fig. 1) on T. cruzi, various concentrations of the peptides were incubated with trypomastigotes. Melittin, the cytolytic and antimicrobial peptide from bee venom, served as a control. After 30 min, viability was recorded by determining the motilities of the trypanosomes by light microscopy. Each peptide used in this study exhibited trypanocidal effects at micromolar concentrations (Table 1). NK-lysin was found to be active against T. cruzi at a concentration of 5 μM, whereas NK-2 and melittin inhibited the motility of trypanosomes at a concentration of 2.5 μM. However, NK-lysin exerted cytopathic and cytolytic effects on the human glioblastoma cell line 86HG39 even at concentrations (i.e., <1 μM) which were tolerated by T. cruzi. In contrast to NK-lysin with its high cytolytic activity, NK-2 was found to be nontoxic for 86HG39 cells. Even in the presence of a 50 μM concentration of NK-2, only minor damage to the human host cells was observed.

FIG. 1.

Position of NK-2 within the structure of NK-lysin. The position of the synthetic peptide NK-2 (dark region) corresponding to amino acid residues 39 to 65 is displayed within the three-dimensional structure of NK-lysin as solved by nuclear magnetic resonance (Brookhaven databank accession no. 1NKL.brk). The primary structure of NK-lysin as deduced from a cDNA sequence (GenBank accession no. CAA59720) is given below the diagram, and the region corresponding to NK-2 is highlighted with an underscore.

TABLE 1.

Inhibition of mobility of T. cruzi by NK-lysin, NK-2, and melittin in comparison to their lytic effects on eukaryotic cells

Peptide	Concn (μM) at which antimicrobial peptides had effects ona:
	T. cruzi	86HG39
NK-lysin	5.0	1.0
NK-2	2.5	>50.0b
Melittin	2.5	2.5

^aThe peptides were incubated with trypomastigotes of T. cruzi or 86HG39 cells. The effect on T. cruzi was visualized by light microscopy. At the indicated concentration, the motility of at least 50% of the parasites had stopped. The effect on 86HG39 cells was determined by PI uptake. At the indicated concentration, at least 50% of the cells were PI positive. Data were determined in three independent experiments, each of which was performed in triplicate.

^bAt 50 μM, only 15% of the cells were PI positive

To examine more directly the effects that these peptides have on T. cruzi, trypomastigotes were incubated with the respective peptides at a concentration of 5 μM, stained with Giemsa, and analyzed by light microscopy (Fig. 2). In a second set of experiments the samples were analyzed by electron microscopy (Fig. 2). Both methods revealed that all peptides used led to morphological alterations. However, the kinds of alterations appeared to be different. Treatment with melittin resulted in total disruption of the cellular membrane (data not shown). By contrast, a swelling of cells was observed after incubation with NK-2 and NK-lysin. Interestingly, incubation with NK-lysin was accompanied by shortening or loss of the flagellum.

FIG. 2.

Morphological changes of T. cruzi trypomastigotes induced by NK-lysin and NK-2. Trypanosomes were incubated with buffer (A; flagellae are marked with an arrow), 5 μM NK-2 (B), or 5 μM NK-lysin (C) and analyzed by light microscopy (Giemsa stain; magnifications of upper panels, ×100) or by scanning electron microscopy (magnifications of lower panels, ×3,000).

To further analyze the mode of action by which the peptides affect T. cruzi, we introduced trypomastigotes expressing β-galactosidase in their cytosol into the study. After incubation of the respective peptide with trypomastigotes for different time intervals, the supernatant was analyzed for β-galactosidase activity by an enzymatic color reaction (Fig. 3). In contrast to the buffer control, all peptides induced the release of β-galactosidase into the supernatant, and the majority of the β-galactosidase was released from the cytosol of T. cruzi as soon as after 15 min.

FIG. 3.

NK-lysin and NK-2 induce release of cytosolic proteins from T. cruzi trypomastigotes. β-Galactosidase-expressing T. cruzi was incubated with 5 μM NK-lysin or NK-2. Melittin (5 μM) and buffer alone served as positive and negative controls, respectively. After the indicated time points, the β-galactosidase activity in the supernatant was analyzed by measuring the absorbance at 560 nm by using CPRG as the substrate. One hundred percent lysis was achieved with 1% NP-40, resulting in an absorbance of 0.16. Experiments were done in triplicate and were repeated twice, with similar results.

As very few trypomastigotes are sufficient to establish an infection by replicating intracellularly in eukaryotic cells, we analyzed their abilities to infect the human glioblastoma cell line 86HG39 after peptide treatment. After incubation of β-galactosidase-expressing T. cruzi with the peptides at different concentrations, the samples were added to monolayers of 86HG39 cells. Five days after infection, the host cells with intracellular parasites were lysed and β-galactosidase activity was determined. Preincubation with each peptide at a concentration of 1 μM substantially diminished the abilities of the parasites to infect cells (Fig. 4). However, if higher concentrations of NK-lysin or melittin were used for preincubation, the transfer of trypanosomes and peptides resulted in lysis of 86HG39 cells, as seen by uptake of PI (Fig. 4, asterisks). This demonstrates again the cytolytic effects of NK-lysin and melittin on mammalian cells, whereas even at higher concentrations NK-2 was nontoxic for these cells.

FIG. 4.

Preincubation of T. cruzi with NK-lysin and NK-2 reduces its infectivity. β-Galactosidase-expressing trypanosomes were incubated with various concentrations of NK-lysin, NK-2, or melittin. Aliquots were removed and used to infect monolayers of the glioblastoma cell line 86HG39. The immediate cytolytic effects of peptides on the host cells are indicated by asterisks. Trypanosomes were allowed to replicate for 5 days and were subsequently analyzed for β-galactosidase activity, which was found to correspond linearly to the T. cruzi number in the culture. Experiments were done in triplicate and were repeated twice, with similar results.

As most trypanosomes reside in the cytosols of their host cells, we examined whether NK-lysin and NK-2 kill intracellular T. cruzi. For this purpose, monolayers of 86HG39 cells were infected with trypomastigotes and extracellular pathogens were removed by washing. Infected monolayers were incubated for 3 or 24 h with different concentrations of the respective peptides, which were tolerated by the host cells. After the indicated incubation period the medium was removed and the remaining cells were washed and cultivated with fresh medium until free trypomastigotes were detectable in control cultures. After this period, the cells were lysed and β-galactosidase activity was analyzed to estimate the amount of T. cruzi in the culture. The lowest concentration of NK-lysin (0.5 μM) which inhibited the growth of trypanosomes was already cytolytic to 86HG39 cells, whereas melittin at 2 μM, a concentration found to be lethal for free trypanosomes, neither affected the growth of intracellular T. cruzi nor induced macroscopically visible damage to the host cell (Fig. 5). In contrast, incubation of infected cells with 2 μM NK-2 diminished the intracellular parasitic load without affecting the host cell. It is interesting that both NK-lysin and NK-2 were highly active in medium supplemented with serum, whereas the lytic effect of melittin, which is known to bind to serum components, appears to be strongly reduced.

FIG. 5.

NK-2 exhibits trypanocidal activity against intracellular T. cruzi. 86HG39 glioblastoma cells infected with β-galactosidase-expressing T. cruzi were incubated with a trypanolytic concentration of NK-lysin, NK-2, or melittin for either 3 or 24 h. The cytolytic activities of the peptides against 86HG39 cells were determined by light microscopy. The cells were washed and cultured in fresh medium until free trypanosomes were visible in buffer-treated cultures. Afterwards, the cells were lysed and subsequently analyzed for β-galactosidase activity. Experiments were done in triplicate and were repeated twice, with similar results.

To corroborate this finding, 86HG39 cells were infected and washed to remove extracellular pathogens. The infected cells were subsequently incubated with sublytic concentrations of the respective peptides (NK-lysin, 0.5 μM; NK-2, 2 μM; melittin, 2 μM) for 24 h, the peptides were then removed, and the cells were further cultivated with fresh medium for another 3 days. Subsequently, the cells were fixed and stained with DAPI for detection of the nuclei and anti-T. cruzi antibodies for detection of the pathogen and counterstained with FITC-labeled phalloidin to visualize host cells. In control cultures, the growth of the trypanosomes was clearly visible, yielding a picture of cells infected with many parasites surrounded by cells harboring only a few intracellular parasites (Fig. 6A). After melittin treatment of the culture, highly parasitized cells were occasionally observed, whereas the number of infected cells in general decreased (Fig. 6B). At higher concentrations melittin was lytic for 86HG39 cells (data not shown). Treatment of infected cultures with NK-2 resulted in a significant decrease in the numbers of infected cells. Interestingly, the highly infected cells, which were frequently seen in control samples, were not found in NK-2-treated samples. Furthermore, the percentage of infected cells in treated cultures was substantially lower than the percentage in the control culture (Fig. 6C). Only a few 86HG39 cells remained in NK-lysin-treated cultures due to the cytotoxic effect of NK-lysin, and the remaining host cells were infected with T. cruzi (Fig. 6D).

FIG. 6.

Effect of peptide treatment on the growth of intracellular T. cruzi. Infected 86HG39 cells were incubated with buffer (A), melittin (B), NK-2 (C), or NK-lysin (D) for 24 h. The cells were subsequently washed and cultured in fresh medium. Cultured cells were fixed and stained with DAPI for detection of nuclei (blue), with FITC-labeled phalloidin for detection of actin (green), and with an anti-T. cruzi antiserum for detection of parasites (red) and observed under a fluorescence microscope. Magnification: ×63. The pictures are representative of two independent experiments.

DISCUSSION

It is present knowledge that the growing class of antimicrobial peptides plays an important role in the innate immune response (18, 37). Most of these peptides have a broad antimicrobial spectrum. Their mode of action, i.e., physical disruption of microbial membranes without the need to interact with a specific receptor, makes the emergence of microbial resistance against peptide antibiotics very unlikely. Persuasive support for this hypothesis comes from the ubiquitous occurrence of such membrane-permeabilizing peptides in nature and their apparent success during the competing evolutions of microorganisms and putative hosts. Collectively, these findings have drawn attention to antimicrobial peptides and to their potential use as novel therapeutic agents against infections (37).

Porcine NK-lysin and its human homolog, granulysin, are produced by lymphocytes, namely, cytotoxic T cells and NK cells, and are effective against a variety of microorganisms in vitro (2, 32). These polypeptides are at least twice as large as most other antimicrobial peptides and appear to be less selective than these, as they are described to be cytolytic to eukaryotic cells as well. NK-lysin exerts potent antimicrobial activity against gram-positive and gram-negative bacteria but also kills the fungal pathogen C. albicans and mammalian cell lines at micromolar concentrations.

Here, we extended the target-cell spectrum to the human parasite T. cruzi, which was found to be susceptible to NK-lysin-mediated lysis. It became apparent that NK-lysin permeabilizes the cytoplasmic membrane of trypanosomes, as indicated by the release of the cytosolic protein β-galactosidase, a marker enzyme introduced into the parasite by DNA transfection. However, the NK-lysin concentration effective against trypanosomes was found to be cytotoxic for mammalian cells, which excludes this therapeutic application of unmodified NK-lysin.

The broad-spectrum activity of NK-lysin has stimulated further attempts to identify the biologically active regions within the polypeptide and to use them as a scaffold to build new synthetic peptides with improved antimicrobial activities and lower toxicities for host cells (3, 4). We have shown that NK-2, the cationic core region of NK-lysin, is substantially less cytotoxic for mammalian cells than NK-lysin but retains its potent antimicrobial activity against bacteria, pathogenic yeast cells, and protozoan parasites (3; this study).

NK-2 is known to permeabilize phospholipid vesicles (3). In good agreement with this previous finding, NK-2 disintegrated the parasitic membrane in such a way that the cytosolic marker protein was released. This indicates that the plasma membrane of the parasite is the target of both peptides, NK-lysin and NK-2. Multiple models which try to illustrate the mechanism by which antimicrobial peptides kill microbes have been proposed. Cell death is most likely accompanied by the formation of pores in the phospholipid bilayer of the microbial membranes. The process of membrane permeabilization may roughly be divided into two distinct mechanisms (25). First, the formation of stable transmembrane pores via the “barrel-stave” mechanism, by which the peptides insert after they bind to the membrane, are then oriented perpendicular to the plane of the bilayer and upon oligomerization create a water-filled channel that is completely lined by the aggregated peptide monomers. Second, the disruption of the membrane via the “carpet” mechanism, by which the peptides bind to the surface of the target until at least a part of it is covered. Destruction of the phospholipid bilayer structure takes place after a threshhold concentration is reached. During this process toroidal pores may be formed transiently, but as the peptides are still on the surface of the bilayer, membrane lesions are lined by both peptides and phospholipids (19). Ultimately, membrane openings brought about by a peptide through any mechanism described so far (regardless of whether such pores are stabile or transient) will affect the internal milieu of the target cell and, if not repaired efficiently, will result in its death.

At present, the consensus view is that only very few membrane-permeabilizing peptides, e.g., alamethicin (5), form stable, barrel-like structured pores. The vast majority, in particular, the cationic and linear alpha-helical antimicrobial peptides such as cecropins and magainins, act by the carpet mechanism (31). Even the cytolytic peptide melittin from bee venom was recently found to induce transmembrane pores that conform to the model of toroidal pore formation (36). NK-2 is a linear amphipathic helical peptide that comprises the region of NK-lysin with the highest density of cationic residues (3). As such, NK-2 is comparable to cecropins and magainins, supporting the likelihood that it also acts by the carpet mechanism.

NK-2 not only was active against extracellular T. cruzi but also inhibited the replication of parasites that reside within the cytosols of host cells. The exact mechanism by which intracellular T. cruzi was killed remains obscure, since the host cells were resistant to NK-2-mediated lysis. One explanation would be that infection by parasites modifies the host cell membrane in a way that exposes a more negative charge and that results in accelerated binding of NK-2 and increased susceptibility of the host cell to peptide-mediated lysis. As in our experiments almost all cells were infected before treatment with NK-2, a selective killing of infected cells that leaves uninfected cells unharmed is unlikely. An alternative explanation would be that only extracellular T. cruzi parasites escaping from the host cells were lysed by NK-2, which prevented secondary infection. However, after infection the number of free trypanosomes remained very small for several days even in controls, and an accelerated release of trypomastigotes from host cells has not been observed during peptide treatment of the culture.

Until now, the comparable activities of peptides against parasites have been shown only rarely. Synthetic peptides derived from the natural membrane-active polypeptides cecropin B and trialysin of insects were capable of lysing extracellular forms of T. cruzi (1, 13). With regard to intracellular parasites, interesting observations have recently been reported for the natural antimicrobial peptide dermaseptin S3 against the protozoan parasite that causes malaria (12). Dermaseptins and some derivatives thereof exhibited potent activities against intraerythrocytic Plasmodium falciparum without disrupting the erythrocyte membrane (15). By using fluorescently labeled peptides, it was shown by videomicroscopy that these dermaseptins bind to the membrane of the infected erythrocytes and subsequently accumulate within the parasite without lysing the host cell. As it appears that the peptide had a higher affinity to the membrane of Plasmodium than to that of the red blood cell, an affinity-driven transfer to the parasite has been proposed (11). Moreover, it has recently been found that highly cationic peptides or arginine- and lysine-rich stretches within larger proteins, i.e., protein transduction domains, may traverse plasma membranes and enter cells without causing toxic effects (20). Such molecular mechanisms may well explain how NK-2 kills intracellular trypanosomes without destroying the 86HG39 host cells.

Collectively, our studies have shown that NK-lysin and its core region, NK-2, are potently trypanocidal. By using the therapeutic index for comparison, the synthetic peptide has properties superior to those of its natural parent molecule. The broad range of antimicrobial activities makes NK-2 a highly interesting lead compound for the development of new anti-infective agents.