Peptide design
Previous studies on the structural characterization of PMAP-23 have shown that PMAP-23 present a random coil structure in aqueous buffer, whereas it forms a helix-hinge-helix structure in membrane mimetic environments. Upon interaction of PMAP-23 with lipid bilayers at the molecular level, the N-terminal helix first binds to negatively charged membrane surface and then inserts the C-terminal helix into the hydrophobic membrane core (Fig. 1). In this study, based on the interaction of PMAP-23 with lipid bilayers, we designed derivatives to make the N-helix more amphiphilic for efficient electrostatic binding to negatively charged membranes and the C-helix more hydrophobic for insertion into hydrophobic interiors. PMAP-N was designed to increase amphiphilicity at the N-terminal helix by exchanging the amino acids between Leu5 and Arg10. To increase the hydrophobic property at the C-terminal helix (PMAP-C), the hydrophilic amino acids (Thr19 and Arg23) were substituted with Lue residues. PMAP-NC was designed to have both increased amphiphilicity at the N-terminus and increased hydrophobicity at the C-terminus. Since the central hinge plays an important role in the interaction of PMAP-23 with lipid bilayers and antimicrobial potency and selectivity [11,13], we retained the central PXXP motif. The sequences and helical wheel diagrams of PMAP-23 and the designed peptides are shown in Fig. 1.
Antimicrobial activity
Table 1 displays the antibacterial activity of PMAP-23 and its rationally designed derivatives as minimal inhibitory concentration (MIC) and minimal bactericidal concentration (MBC) against gram-positive bacteria, including S. aureus and B. subtilis, as well as gram-negative bacteria, including E. coli and P. aeruginosa. PMAP-23 exhibited activity with MIC values in the 4–8 µM range and MBC values in the 16–64 µM range against the tested strains of bacteria. Importantly PMAP-NC was found to be the most potent amongst the peptides against both gram-positive and gram–negative bacteria, with MIC values in the 2 to 4 µM range and MBC values in the 4 to 8 µM range. The MIC values between PMAP-N and PMAP-C were similar or showed only a two-fold difference. In MBC values, however, PMAP-C exerted bactericidal activity about 4 times better than PMAP-N, suggesting that the hydrophobic C-terminal helix is crucial for the bactericidal activity.
Table 1
Minimal inhibitory concentration (MIC) and minimal bactericidal concentration (MBC) of peptides against S. aureus, B. subtilis, E. coli and P. aeruginosa.
|
Antimicrobial activity (µM)
|
peptide
|
S. aureus
|
B. subtilis
|
E. coli
|
P. aeruginosa
|
MIC
|
MBC
|
MIC
|
MBC
|
MIC
|
MBC
|
MIC
|
MBC
|
PMAP-23
|
4
|
16
|
4
|
16
|
4
|
32
|
8
|
64
|
PMAP-N
|
2
|
16
|
4
|
16
|
4
|
32
|
8
|
64
|
PMAP-C
|
2
|
4
|
4
|
8
|
4
|
8
|
8
|
16
|
PMAP-NC
|
2
|
4
|
2
|
4
|
4
|
8
|
4
|
8
|
To further investigate the differences in their lethal effects, we measured the time-dependent behavior of their bactericidal activity against S. aureus and E. coli, which are examples of gram-positive and gram-negative bacteria, respectively. The number (CFUs/ml) of viable bacterial cells trended to decrease with time, after treatment with peptides to the bacteria (Fig. 2). At the 2 x MIC, PMAP-23 and PMAP-N had little immediate impact on the number of surviving CFUs which were reduced by less than only 20% after 120 min of incubation. By contrast, it was observed that PMAP-C and PMAP-NC almost completely killed both bacteria within 120 min. Killing kinetics studies of the peptides on S. aureus and E. coli demonstrated that the hydrophobic C-terminal α-helix has a drastic effect on the rapid killing of the peptides.
Membrane depolarization and permeabilization
The bactericidal activity of AMPs is often associated with membrane depolarization and permeabilization. To investigate whether bacterial killing by PMAP-23 and its derivatives is related to membrane depolarization, the ability to depolarize the cell membrane of S. aureus was evaluated (Fig. 3a). All peptides dissipated the membrane potential in a dose-dependent manner. Interestingly, we found that PMAP-C and PMAP-NC were much higher than PMAP-23 and PMAP-N, indicating that the hydrophobic C-terminal helix has a significant effect on their membrane depolarizing activity. The correlation between bactericidal activity and membrane depolarization of the peptides suggests that loss of membrane potential may be a major contributor to the bactericidal effect. We then generated calcein-containing liposomes composed of PG/PC (1:1) and investigated the ability of the peptides to induce membrane permeabilization by measuring the release of entrapped calcein from the liposomes into an aqueous buffer (Fig. 3b). Similar to their ability to depolarize bacterial membranes, PMAP-23 and PMAP-N induced weak calcein release (~ 30%) from liposomes, whereas PMAP-C and PMAP-NC showed strong membrane-lytic activity (70 ~ 80%), which further confirmed that the C-terminal helix plays important roles in efficient membrane permeabilization. PMAP-NC with the most potent bactericidal activity showed the greatest membrane depolarization and permeabilization, suggesting that membrane disruption is the primary event in bacterial killing for PMAP-NC. The potential intracellular targets associated with the antimicrobial activity of PMAP-23 led us to investigate the peptides' ability to translocate across PC/PG (1:1) liposomes. We accomplished this by monitoring resonance energy transfer from the Trp residues of PAMP-23 and its derivatives to the dansyl-PE, which was incorporated into the liposomes (Fig. 3c). We observed that all peptide efficiently translocated across the lipid bilayers but they exhibited only slight variations in their capacity to translocate across the membrane, indicating that the translocation does not play decisive role in the bactericidal efficacy of the peptides.
Cytotoxicity of PMAP-23 and PMAP-NC against human cancer cells and erythrocytes
We evaluated the anticancer potential of PMAP-23 and PMAP-NC on human cancer cells (MDA361 and human A549) using the MTT assay, and their selectivity was assessed using human erythrocytes. While PMAP-23 was almost inactive, PMAP-NC demonstrated robust anticancer activity against MDA361 and A549 cancer cell lines (Fig. 4a and 4b). Even though PMAP-NC at high concentration (64 µM) elicited 20% release from human erythrocytes, the peptide was considerably more toxic to cancer cells than to human erythrocytes (Fig. 4c). As cancer cells display exposed negatively charged PS on their surface, we further evaluated the potential contributions of PS in peptide-induced membrane permeability. The peptides exhibited a comparatively low ability to disturb zwitterionic PC liposomes, which is consistent with the results obtained for their hemolytic activity. Interestingly we found that consistent with the data from cell-based system, PMAP-23 did not have a noticeable impact on the permeabilization of PS-containing liposomes, whereas PMAP-NC exhibited an increase in permeabilization activity with a rise in PS content (Fig. 4d). These results can potentially clarify why PMAP-NC has selective anticancer properties and also indicate that PMAP-NC may serve as a lead compound for developing anticancer peptides.
Partitioning of peptides into lipid bilayers by Trp fluorescence analysis
All peptides presented in Fig. 1 have two Trp residues, one each on the N- and C-helices. Because the intensity and emission maximum of Trp fluorescence are highly responsive to changes in the environment surrounding Trp residues, we utilized Trp fluorescence properties to estimate the partition of peptides into lipid membranes. Table 2 shows the Trp fluorescence emission maxima (λmax) and Stern-Volmer constants (KSV) for PMAP-23 and its derivatives in buffer and in the presence of vesicles composed of PC/PG (1:1), PC/PS (9:1), or PC, at a peptide/lipid molar ratio of 1:100. The KSV values were determined by conducting a fluorescence quenching experiment in which a water-soluble fluorescence quenching agent called acrylamide was used. The wavelength maxima range of the peptides in aqueous buffer was 352–353 nm, indicating that Trp residues were hydrophilic in nature. In the presence of PC/PG (1:1) vesicles, all peptides caused a blue shift (10–11 nm) in the emission maximum, indicating that they bind to the negatively charged membranes. In contrast, the addition of peptides to PC liposomes only caused a minor shift (1–3 nm) in the Trp emission maxima. This is due to the low exposure to acrylamide when Trp residues are buried in the lipid bilayer, leading to a low KSV value. In comparison to PC vesicles, the KSV values of PMAP-23 and its derivatives were lower in PC/PG (1:1), which suggests that the Trp residues of peptides penetrate deeper into the hydrophobic core of lipid membranes. These findings explain why the peptides have a selective antimicrobial effect due to the fact that eukaryotic membranes are mostly made up of zwitterionic phospholipids, while prokaryotic membranes are composed of a combination of negatively charged and zwitterionic phospholipids. Interestingly, in the presence of PC/PS (9:1) vesicles, the Trp emission maxima of PMAP-C and PMAP-NC were more shifted than those of PMAP-23 and PMAP-N, suggesting that the increased hydrophobicity at the C-terminus is essential for the strong and deep insertion into the hydrophobic core of the cancer-mimicking membrane. Importantly, the lower KSV value of PMAP-NC over PMAP-23 in PC/PS (9:1) vesicles further explains the direct correlation between the binding of peptides to cancer-mimicking membranes and their superior anticancer activity.
Table 2
Tryptophan fluorescence emission maxima (λmax) and Stern-Volmer constants (KSV) for PMAP-23 and its derivatives in buffer and in the presence of vesicles composed of PC/PG (1:1), PC/PS (9:1), or PC at a peptide/lipid molar ratio of 1:100.
peptide
|
Buffer
|
PC/PG (1:1)
|
PC/PS (9:1)
|
PC
|
λmax
|
KSV
|
λmax
|
KSV
|
λmax
|
KSV
|
λmax
|
KSV
|
PMAP-23
|
352
|
15.6
|
343
|
7.5
|
349
|
8.3
|
351
|
14.7
|
PMAP-N
|
352
|
15.3
|
342
|
7.8
|
348
|
8.2
|
351
|
14.2
|
PMAP-C
|
353
|
15.8
|
342
|
6.3
|
344
|
6.6
|
350
|
14.5
|
PMAP-NC
|
353
|
15.5
|
342
|
5.5
|
344
|
5.8
|
350
|
14.3
|
Comparison of structural and functional properties between PMAP-23 and PMAP-NC
In the design and development of novel AMPs for the treatment of antibiotic-resistant bacteria, there has been considerable interest in understanding the structure-activity relationship of AMPs and their interactions with lipid bilayers. In general, broad-spectrum antibiotic activities are typically the result of the binding of antimicrobial peptides to target cell membranes, as well as the disruption of these membranes. This binding and disruption are primarily due to the electrostatic and hydrophobic interactions that occur between the peptides and the bacterial membranes (Wimley 2010; Lee et al. 2016). PMAP-23 adopts amphipathic helix, hinge, and amphipathic helix. The initial binding of PMAP-23 to the anionic membrane surface is mainly due to the N-terminal amphipathic helix. On the other hand, the efficient and rapid insertion of the peptide into the hydrophobic core of the membrane largely depends on the C-terminal hydrophobic helix. The central hinge region, which is an important structural component, enables a separation between the N-terminal electrostatic interaction and the C-terminal hydrophobic interaction with the target cell membranes. Based on the interaction of PMAP-23 with membranes, therefore, we designed PMAP-NC to have a more amphipathic at N-terminus and more hydrophobic at C-terminus.
In bacterial growth inhibition tests, PMAP-23 and PMAP-NC acted similarly against gram-positive S. aureus and gram-negative E. coli. In bacterial killing assays, however, PMAP-NC killed both S. aureus and E. coli rapidly, whereas the same concentration of PMAP-23 was negligible even after 120 min of contact with the bacterial strains (Fig. 2). Compared with PMAP-23, PMAP-NC dissipated membrane potential to a greater extent in S. aureus (Figs. 3a). Likewise, PMAP-NC induced greater leakage of calcein from LUVs than PMAP-23 (Fig. 3b). Considering that PMAP-C lead to efficient membrane depolarization and permeabilization and exerted more improved the bactericidal activity than PMAP-N, the insertion of hydrophobic C-terminal helix into the membrane is more critical step than the binding of amphipathic N-terminal helix to the membrane in the process of bactericidal activity. Therefore, the rationally designed PMAP-NC showed the strongest and fastest bactericidal activity due to efficient membrane disruption.
In recent years, AMPs have garnered extensive attention for their potential as a new class of anticancer drugs to overcome tumor resistance to conventional chemotherapy (Gaspar et al. 2013; Jafari et al. 2022; Deslouches and Di 2017). While AMPs are believed to interact with the membrane surface of cancer cells, disrupting their membranes and causing cancer cell death, the structural and physicochemical parameters that determine their anticancer activity are still poorly understood. In order to design anticancer peptides without cytotoxic activity against normal cells, great efforts are being made to understand the difference between normal cells and cancer cells. We showed that PMAP-NC significantly reduced cell viability of cancer cells with low cytotoxicity against human erythrocytes. The membrane permeability of PMAP-NC was dependent on the PS content, which explains the selective anticancer activity. The unique structure of PMAP23, which forms an amphiphilic N-helix and a hydrophobic C-helix connected by a central hinge, plays a key role in performing selective membrane disruption of PS-containing liposomes. These results may provide a good strategy for developing selective anticancer peptides.
In conclusion, PMAP-23 showed only antibacterial activity whereas PMAP-NC were toxic to bacterial and cancer cells. PMAP-23 required a delay time for its bactericidal effect, but PMAP-NC was able to kill bacteria rapidly. The fast kinetics of bactericidal activity caused by PMAP-NC are linked to efficient membrane depolarization and permeability, indicating that membrane disruption is the key event in bacterial killing for PMAP-NC. Comparison of PMAP-N and PMAP-C demonstrated that the hydrophobic C-terminal α-helix was more important for the membrane disruption ability of PMAP-NC than the amphipathic N-terminal α-helix. Importantly we found that PMAP-NC disrupted PS-containing liposomes as a function of PS concentration, which explains its relatively strong anticancer activity. Taken together, our results suggest that a unique structure with an amphipathic N-helix, a central hinge, and a hydrophobic C-helix is responsible for bactericidal and anticancer activity of PMAP-NC. We propose that detailed knowledge of the interaction of peptides with various helix-hinge-helix structures with lipid bilayers may facilitate the design of highly selective and potent AMPs as effective antimicrobial and/or anticancer drugs in the future.