Anti-plasmodial action of de novo-designed, cationic, lysine-branched, amphipathic, helical peptides

The devastating diseases caused by protozoan parasites are a major burden of the tropics, and in particular, Plasmodium falciparum, the causative agent of falciparum malaria, creates a serious public health problem in many areas of the densely populated developing world. The widespread resistance of P. falciparum to chloroquine (CQ), which has spread from Asia to Africa, has rendered the drug ineffective against the most dangerous Plasmodium strain in many affected regions of the world. Unfortunately, CQ-resistance is associated with cross-resistance to other quinoline drugs, such as quinine and amodiaquine [1]. Plasmodium falciparum is genetically diverse and has multiple independent origins of mutations in genes that confer resistance to widely used anti-malarial drugs [2]. Left with just artemisinin to fight against malaria, Arata Kochi, Director of the Malaria Division at the World Health Organization, had felt compelled to say “if we lose artemisinin, we will no longer have an effective cure for malaria” [3]. However, most recently, alarming signs of clinical resistance against artemisinin, in the form of delayed parasite clearance, are being observed in the border between Cambodia and Thailand [4, 5]. The challenge of designing an effective vaccine along traditional lines against malaria is that many P. falciparum proteins are highly polymorphic and their functions are redundant [6]. More than 200 million new malaria cases reported annually is a challenge [7] that underscores the urgent requirement for new drugs against malaria.

Peptides are an essential component of defence mechanism of all life forms and anti-microbial peptides are evolutionarily ancient biological weapons. Their widespread distribution throughout the living kingdom suggests that anti-microbial peptides may have served a fundamental role in the successful evolution of complex multi-cellular organisms [8]. Despite their ancient lineage, anti-microbial peptides have remained effective defensive weapons, defeating the general belief that bacteria, fungi and viruses can and will develop resistance to any conceivable substance. Among other differences, uniquely anionic charge on bacterial surface is a curious feature that distinguishes the prokaryotic bacteria from their eukaryotic counterparts [9]. Anti-microbial peptides gain selectivity from their ability to target this previously under-appreciated ‘microbial Achille’s heel’ [10‐12]. Interestingly, a seminal feature of the malaria parasite-infected red cell is reflected in an altered asymmetry of lipid composition in its cell surface membrane. In contrast to the uninfected, healthy red cell, the malaria-infected red cell shows a translocation of the anionic phosphatidylserine from the inner leaflet to the outer leaflet of the bi-layer [13]. As a result, the FITC-Annexin negative, healthy red cell now turns to become FITC-Annexin positive [14]. Thus a Plasmodium-infected red cell seems to mimic the anionic surface charge that characterizes the bacterial cell surface. In principle, this is expected to make malaria-infected red cells become vulnerable to the action of anti-microbial peptides. Indeed, naturally occurring or modified peptides, such as dermaseptin [15], oligoacyllysine [16], cyclosporin A [17], cecropin A [18], NK-2 [14] and meucin [19], have been found to display in vitro anti-malarial activity. Some membrane-active, hydrophobic peptides of fungal origin have also been found to exhibit in vitro anti-malarial action [20]. However, many of these naturally occurring peptides suffer from drawbacks such as poor potency, stability and selectivity [21]. Therefore, in a bid to improve their performance, efforts are being made to engineer peptides in diverse ways with the aim of reducing their size, improving their stability against proteases and enhancing their selectivity [22‐24]. The structure activity relationships of a series of de novo-designed, conformationally-constrained helical, amphipathic, cationic peptides against bacteria have earlier been reported [25]. In the present work, the potent anti-plasmodial action of these peptides against both CQ-sensitive and CQ-resistant strains of P. falciparum are being reported. The results indicate that a lysine-branched, dimeric peptide ΔFd, which is highly potent (IC₅₀ 1.39 μM) across CQ-sensitive and CQ-resistant strains {Resistance Index (IC₅₀ CQ resistant strain/ IC₅₀ CQ sensitive strain)1.1} of P. falciparum, fairly selective against parasitized red blood cells {Selectivity Index (HC₅₀ URBC/IC₅₀P. falciparum) >35) and fairly non toxic to mammalian HeLa cells (TC₅₀ >25 μM), stalls parasite growth by causing arrest of ring stage parasite, anomalous egress of trophozoites and premature egress of schizonts that fail to produce invasion competent merozoites.

The success of antibiotics is based upon the characteristic molecular targets that distinguish the prokaryotic bacteria from the nucleated eukaryotic cells [32, 33]. Cationic, amphipathic helical, antibiotic peptides also seem to gain specificity by exploiting the fact that bacteria have a preponderance of anionic lipids, such as phosphatidylglycerol and bis(phosphatidyl)glycerol (cardiolipin), conferring a negative charge on their surface. In contrast, their eukaryotic counterparts have a high density of zwitterionic lipids such as phosphatidylcholine and phosphatidylethanolamine, enabling their surfaces to be largely neutral [34, 35]. A well-studied and yet curious feature of the human red blood cell is the transition from FITC-Annexin negative to FITC-Annexin positive status upon infection with the malaria parasite [14]. It is well known that this phenomenon is caused by the translocation of the anionic phosphatidylserine from the inner to the outer leaflet of the lipid bilayer. Thus infection with Plasmodium confers an anionic character to the red blood cell giving it a shade of semblance to a bacterial membrane. Focusing on the altered membrane asymmetry seen in the infected red cell, the interesting anti-plasmodial properties of several de novo-designed, cationic, amphipathic, helical, bonafide membrane-active anti-bacterial peptides have been examined in the present studies.

The first observation of the comparative anti-plasmodial potencies of two monomeric (ΔFm, Fm) and four dimeric peptides {ΔFd, Fd, D-Lys-ΔFd, (ΔFm)₂}(Table 1) indicated that the dimers (IC₅₀ 1.39- 2.4 μM) were about two orders of magnitude more potent than the monomers (IC₅₀ >100 μM). Among the dimeric peptides {ΔFd: IC₅₀ 1.39 μM, D-Lys-ΔFd: IC₅₀ 1.8 μM, (ΔFm)₂: IC₅₀ 2.4 μM, Fd: IC₅₀ 1.5 μM}, the lysine-branched ΔFd was chosen for detailed mechanistic studies since it had favourable features of anti-plasmodial potency and selectivity index (>35). Interestingly, in going from this bivalent-branched dimer ΔFd to the tetravalent K-K₂-branched quadrumer ΔFq (IC₅₀: 0.25 μM), a further six-fold potentiation was observed. However, as shown in Table 1, this potentiation was associated with a decline in selectivity index from >35 (ΔFd) to 10 (ΔFq). Nevertheless, the trend of increasing anti-plasmodial potency with increasing valency (monomer to dimer to quadrumer) of the peptide suggests that oligomerization on cell surfaces may play an important role in the anti-plasmodial action of these cationic, amphipathic peptides. It is important to note that crystal structures of several ΔF-containing peptides have revealed the propensity of this planar aromatic residue to engage in long-range, multicentred interactions (N-H…O, C-H…O, C-H…π, and N-H…π) that can stabilize oligomeric states like the ΔF zipper [36] in the absence of linker, or the helical hairpins in the presence of appropriate linker [37, 38]. The coming together of optimal values of anti-plasmodial potency and selectivity indices (haemolytic selectivity index >35 and mammalian cell cytotoxicity index >16) in the lysine-branched dimeric ΔFd became the motivation to unravel mechanistic details of the anti-plasmodial action of this peptide. (ΔFm)_2, the corresponding linear dimer, was also studied in some experiments to explore if the mode of dimerization may influence the anti-plasmodial actions of these two dimeric peptides.

The essentiality of apicoplast, an organelle of cyano-bacterial origin in the malaria parasite, is well known to make the parasite vulnerable to antibiotics,such as tetracycline, clindamycin and thiostrepton [39, 40], which are known to cause delayed death in malaria parasite. This phenomenon, caused by targeting of the apicoplast or the mitochondrion, is characterized by a significantly lower IC₅₀ post second cycle (at 96 hr) vs the first cycle (at 48 hr) [41]. In order to find if the anti-plasmodial peptides under study may be targeting organelles such as the apicoplast of the malaria parasite, comparative anti-plasmodial potencies against P. falciparum 3D7 were determined at both 48 hr and 96 hr. Since the data (Figure 1) did not show a significant reduction of IC₅₀ at 96 hr, the possibility that ΔFd and (ΔFm)₂ may cause early death by influencing several other targets besides the apicoplast and the mitochondria cannot be ruled out.

In trying to gain a better understanding of the probable mechanisms that confer the malaria parasite growth inhibitory properties on the dimeric peptide ΔFd, the peptide-treated samples were examined by microscopy. As shown (Figure 2A), in comparison to the untreated control (high ring-stage parasitaemia), while the IC₅₀-treated ring stage synchronized culture was found to have the initial low parasitaemia (1%) and growth arrest at trophozoite stage, the IC₈₀ treated sample was found to have the intial parasitaemia (1%) with the few parasitized cells showing arrested, probably dead pyknotic ring forms. Interestingly, the microscopic examination of the ΔFd-treated, trophozoite-enriched culture (Figure 2B) showed the presence of extra erythrocytic Giemsa-positive trophozoites alongside uninfected red blood cells. Indeed manual counting of a large number of fields (Additional file 8) indicated that over 95% of the trophozoites were in fact extracellular. The presence of extracellular trophozoites in the midst of intact uninfected red blood cells was suggestive of the selective haemolytic action of ΔFd on parasitized cells causing anomalous release of trophozoites following 24-hr incubation.

The transition of ring stage to trophozoite stage in presence of ΔFd at its IC_50, indicates that while this low dose is sufficient to arrest trophozoites, it is clearly not sufficient to halt the ring from moving to the trophozoite stage (Figure 2A). This heightened sensitivity of trophozoite-stage cultures over ring-stage cultures may be related to the enormous red cell reorganizational changes associated with a fast feeding, actively metabolizing and replicating life style of trophozoite in comparison with the more sedentary ring stage. The selective lysis of trophozoite-bearing cells (Figure 4B) also suggests that such remodelling of the trophozoite harbouring red cell membrane [42] may be rendering it more vulnerable to the action of membrane active peptides like the ΔFd. The greater vulnerability of trophozoite bearing over ring-bearing red cells is evident also from the fact that peptide-mediated haemolysis is directly proportional to the percentage trophozoites in mixed stage culture samples (Figure 4A).

Trophozoite egress, induced by the peptide, is not natural to the life cycle of the malaria parasite. Hence, it was important to find if host cell membranes-may be involved in the process. To address this issue, the peptide-treated sample was exposed to immunostaining with band 3 antibody. As shown (Figure 5), the egressed trophozoites were not flanked by band 3 staining suggesting that the process is more likely to be caused by lysis of the infected host cell. A closer examination of the phenomenon of ΔFd-mediated anomalous egress of trophozoites revealed that in trophozoite-ring mixed culture exposed to FITC-ΔFd at low concentrations (2 μM) and short time (30 min) (Figure 5A), the peptide seems to enter and attack the parasite from within without causing immediate lysis of the infected red cell. However when trophozoite-stage culture was exposed to ΔFd for longer times (24 hr), at similar low concentrations (2 μM), this peptide seemed to cause selective lysis of parasitized cells leading to anomalous egress of trophozoites (Figure 2B). Further, at higher concentrations (12.0 μM) this peptide caused selective lysis of red cells harbouring trophozoites within 3 hr (Figure 4B).

Unlike trophozoites, schizonts have an intrinsic program of egress that causes the release of numerous merozoites leading to infection of fresh red cells causing amplification of infection and increasing the severity of disease. Hence it was interesting to find if ΔFd may perturb the programmed process of egress in schizonts. As shown (Figure 2C and Additional file 10), this peptide caused premature egress of schizonts. As a consequence, the egressed schizonts which showed lumps of amplified DNA did not exhibit the characteristic symmetrically organized rosette appearance of merozoites seen in the untreated control schizont. Further the merozoites from the peptide treated culture showed a significantly low invasion efficiency in comparison to the control merozoites (Additional file 10). Figure 8 summarizes the versatility of ΔFd to target each stage of the life cycle of P. falciparum in characteristic and decisive ways with good selectivity.

Malaria parasites go to extraordinary means to modify RBC membrane, which separates them from the external world. These modifications include a marked increase in erythrocyte membrane fluidity [43‐46], alterations in host cell lipid fatty acid composition [47, 48] and phospholipid-transbilayer distribution [49], enhancement of the rate of lipid transbilayer movement [50, 51] and increased permeability through newly formed pores on the erythrocyte membranes [52, 53]. As a part of these major re-organizational events, the malaria-infected red cell is well known to exhibit a translocation of the anionic phosphatidylserine from the inner leaflet to the outer leaflet of the bi-layer [13, 54]. This more negative cell surface may provide the force for the fast and specific uptake of cationic peptides by the malaria-infected red cell. Previous studies have indicated that high levels of cellular uptake can be achieved through the inclusion of cationic residues into arginine-based peptide oligomers [55]. The positive molecular charge facilitates charge-driven uptake through the plasma membrane, which exhibits a potential gradient that can electrophorese cationic species from the extracellular space into the cell [56, 57]. Interestingly, a recent study has demonstrated that membrane asymmetry can be altered and maintained in the altered state by externally added poly-L-lysine [58]. The combined microscopic (Figure 6) and FACS analysis (Figure 7) suggests that ΔFd enters the infected cells and stains rings and trophozoites within a few minutes. Thus it is quite likely that ΔFd and (ΔFm)₂, the two cationic dimeric peptides studied here, in close resemblance to poly-L-lysine, may first home on those infected red blood cells that show slightly more anionic character as a result of alterations in membrane asymmetry and binding of these cationic peptides could further enhance and maintain this anionic character facilitating the stronger binding and faster internalization of peptides into the infected cells.

The ability of ΔFd to cross the host red cell membrane, the parasitophorous vacuole membrane, the parasite plasma membrane and also the parasite nuclear membrane to reach the nucleus of the parasite (Figure 6B), indicates its resemblance to cell-penetrating peptides which are known to have a lipophilic-cationic character. Even as the peptide was apparently targeting the DNA of the parasite, the absence of FITC-ΔFd on the host red cell membrane or all the subsequent membranes mentioned above was puzzling. It was surmised that these localizations may have been missed due to the rapidity of the process of peptide uptake. In order to capture some stages preceding the intranuclear entry of the peptide, the peptide-staining experiment was performed as a function of both time and temperature. As shown (Figure 6C), the images captured at 4°C (30 min) indeed showed predominant staining on the host cell surface. In contrast, the images corresponding to 25°C (10 min) and 25°C (30 min) showed progressively greater staining of the intracellular parasite nuclear material. These results suggest that this peptide crosses several membranes of the infected red cells before entering the nucleus.

The most probable reasons for the significantly enhanced potency of the dimers ΔFd/Fd over the monomers ΔFm/Fm include increased membrane binding and permeabilization, enhanced binding affinity for DNA and proteins and enhanced biochemical stability against degradation by proteases. These properties originating from increased avidity and affinity of interactions unique to dimeric peptides and absent in monomeric peptides have been described previously [25]. In studies on the antibiotic action of these peptides it has previously been observed that the requirements of helicity for potent antibiotic action are much higher for the gram positive Staphylococcus.aureus than is the case with the Gram-negative Escherichia coli. In contrast, as shown in the present study, all dimers {(ΔFd, Fd, D-Lys-ΔFd, (ΔFm)₂} are nearly equipotent against P. falciparum (Table 1). This suggests that different conformational and topological properties of peptides may be important for their activity against different organisms.

Some important features of these peptides as drugs against malaria include their favourable resistance indices (Table 1) that allow them to rapidly kill both drug-sensitive and drug-resistant strains of malaria parasite with equal potencies, their amphipathic nature that gives them drug-like character, and their ability to permeabilize and penetrate biological membranes, which allows them to attack target cells both from the surface as well as intracellularly. In addition, the presence of the conformationally constrained, non-protein, amino acid didehydrophenylalanine in both ΔFd and (ΔFm)₂ provides considerable protection against proteolytic degradation [25]. Even as these two dimeric peptides offer similar profiles of anti-plasmodial actions, a judicious choice for further improvisation should be the branched dimer ΔFd over the linear dimer (ΔFm)₂ since (a) the former is little more potent against P. falciparum, (b) the branched dimer is more stable against proteases [25], and (c) the branched dimer has better economics of production since the time it takes to synthesize a branched dimer is half as much as the time it takes to assemble a linear dimer.

This study reports the anti-plasmodial action of ΔFd, a de novo-designed, cationic, lysine-branched amphipathic, helical peptide. In vitro assays suggest good selectivity (>35), good resistance index (1.1) and low mamamalian cell cytotoxicity, as a promise of ΔFd against malaria. The strategy adopted by ΔFd to inhibit the growth of malaria parasite appears to be broadly two-fold: (a) involving growth arrest without causing lysis of red cell (at IC₅₀-IC₁₀₀), and (b) anomalous egress of trophozoites and premature egress of undifferentiated schizonts leading to death of the parasite (at > IC₁₀₀).