Deciphering Structure-Function Relationship Unveils Salt-Resistant Mode of Action of a Potent MRSA-Inhibiting Antimicrobial Peptide, RR14

ABSTRACT

Multidrug-resistant (MDR) bacteria lead to considerable morbidity and mortality, threatening public health worldwide. In particular, infections of methicillin-resistant Staphylococcus aureus (MRSA) in hospital and community settings are becoming a serious health problem. Antimicrobial peptides (AMPs) are considered novel therapeutic targets against MDR bacteria. However, salt sensitivity reduces the bactericidal potency of AMPs, posing a major obstacle for their development as antibiotics. Thus, the design and development of salt-insensitive peptides with potent antibacterial activity is imperative. Here, we employed biochemical and biophysical examinations coupled with molecular modeling to systematically investigate the structure-function relationship of a novel salt-insensitive AMP, RR14. The secondary structure of RR14 was characterized as an apparent α-helix, a structure that confers strong membrane-permeabilizing ability targeting bacterial-mimetic membranes. Additionally, the bioactive structure of RR14 was determined in complex with dodecylphosphocholine (DPC) micelles, where it possesses a central α-helical segment comprising residues R4 to K13 (R4–K13). RR14 was observed to orient itself into the DPC micelle with its N terminus and the α-helical segment (I5–R10) buried inside the micelles, which is essential for membrane permeabilization and bactericidal activity. Moreover, the specific and featured arrangement of positively charged residues of RR14 on its amphipathic helical conformation has great potential to render its strong salt resistance ability. Our study explored the structure-function relationship of RR14, explaining its possible mode of action against MRSA and other microbes. The insights obtained are of great applicability for the development of new antibacterial agents.

IMPORTANCE Many antimicrobial peptides have been observed to become inactive in the presence of high salt concentrations. To further develop new and novel AMPs with potent bactericidal activity and salt insensitivity, understanding the structural basis for salt resistance is important. Here, we employed biochemical and biophysical examinations to systematically investigate the structure-function relationship of a novel salt-insensitive AMP, RR14. RR14 was observed to orient itself into DPC micelles with the N terminus and the α-helical segment (I5–R10) buried inside the micelles, which is essential for membrane permeabilization and bactericidal activity. Moreover, the specific and featured arrangement of cationic residues of RR14 on its amphipathic helical conformation renders its strong salt resistance ability. The insights obtained are of great applicability for developing new antibacterial agents.

INTRODUCTION

Drug-resistant bacteria, rapidly emerging to exert considerable pressure on the use of conventional antibiotics, further threaten the health of people all around the world (1). In particular, infections of methicillin-resistant Staphylococcus aureus (MRSA) lead to severe health problems in hospital and community settings worldwide (2–4). Of note, soft tissue and skin infection epidemics were observed to be mainly caused by community-associated MRSA worldwide since the late 1990s (5, 6). Also, the mortality rates resulting from MRSA infections have been estimated to be higher than those of tuberculosis and HIV infections in U.S. hospitals (7). Even worse, the resistance of S. aureus to multiple antibiotics creates a considerable challenge in the treatment of MRSA infections and limits other therapeutic options (8–10). Therefore, the development of new bactericidal agents with novel modes of action and without the likelihood of evolving resistance are urgently needed to combat MRSA.

Most creatures are endowed with the ability to defend against infections by microorganisms. In nature, organisms evolutionarily obtain defensive abilities from progenitors to recognize and resist attacks by pathogens. In particular, antimicrobial peptides (AMPs) are key components of the innate immune system observed in most living organisms (11–13). AMPs function by binding to and disrupting bacterial membranes and can kill a broad spectrum of drug-resistant microbes with MICs in the micromolar range (14–16). Additionally, the membrane-disrupting AMPs can avoid the evolved resistance of bacteria due to the conservation of lipid compositions critical for the construction of bacterial membranes (17, 18). Therefore, AMPs are of great potential for the development of therapeutics against antibiotic-resistant bacteria (19–21). Although AMPs are potent in inhibiting microbes, they have several limitations as therapeutic agents. Generally, natural AMPs often exist in long lengths (high cost for production) and have moderate bactericidal activity and apparent cytotoxicity (22–24). Besides, most natural AMPs show significantly decreased antibacterial ability when exposed to physiological salt conditions (25, 26). This drawback further impedes the medical development and application of AMPs to be feasible as therapeutic agents. Accordingly, short peptides (either natural or synthetic) with potent bactericidal activity, low cytotoxicity, and high salt tolerance are ideal candidates for the development of new AMPs for novel therapeutic agents.

Several short synthetic peptides have been developed for therapeutic applications against microbial infections (27–33). Recently, a synthetic peptide, RR14 (RRIKA), was reported to exert potent bactericidal activity against clinical and drug-resistant Staphylococcus strains (34). The AMP, RR14, was observed to inhibit the growth of linezolid-resistant S. aureus, methicillin-resistant Staphylococcus epidermidis, methicillin-sensitive S. aureus (MSSA), MRSA, vancomycin-resistant S. aureus (VRSA), and vancomycin-intermediate S. aureus (VISA) with MICs of 2 to 4 μM (34). Remarkably, RR14 retained its antibacterial activity in the presence of physiological salt concentrations—the MICs of RR14 against S. aureus strains did not increase as the concentrations of NaCl and MgCl₂ increased to 100 and 2 mM, respectively (34). Moreover, RR14 was found to act synergistically with lysostaphin to eradicate the growth of VRSA, MSSA, and MRSA (34). Although the salt tolerance property of RR14 is evident, the detailed structure-function relationship associated with its mode of action has not been investigated. Mohamed et al. reported that RR14 may interfere with and disrupt the integrity of the bacterial membrane, resulting in the leakage of cytoplasmic contents, followed by cell death (34). However, the interaction of RR14 with bacterial membrane (lipids) was not thoroughly elucidated. Usually, progress in developing AMPs as potent bactericidal agents is hindered by a poor understanding of the structure-function relationship, which correlates highly with its mechanism of action. To further develop new and novel AMPs with potent bactericidal activity and salt insensitivity, understanding the detailed mode of action of RR14 and exploring the key residues contributing to its binding with bacterial membranes are imperative.

Therefore, to thoroughly understand the bactericidal mode of action of RR14, we aimed to investigate and characterize its structural properties and membrane-disrupting ability, as well as the structure and orientation upon binding to membrane-mimetic micelles in solution. In this study, we employed biophysical and biochemical methodologies combined with molecular modeling to explore the structure-function relationship of RR14 targeting dodecylphosphocholine (DPC) micelles. The secondary structure contents, membrane-permeabilizing ability, micelle-bound solution structure, orientation in the DPC micelles, and modeled complex structure of RR14-micelles were investigated and determined by circular dichroism (CD) spectroscopy, calcein leakage assay, solution state nuclear magnetic resonance (NMR) spectroscopy, spin label-enhanced NMR spectroscopy, and structural modeling, respectively. The structure-function insights obtained could explain the mode of action of RR14 against MRSA and other microbes. Our results provide valuable information that will be of benefit for the development of new and novel AMPs to combat MRSA infections.

RESULTS

The structural properties of RR14.

To explore the structure-function relationship of RR14, we first investigated the secondary structural properties by using far-UV CD spectroscopy. The CD spectra of RR14 demonstrated a random coil conformation in aqueous solvent (20 mM sodium phosphate, 100 mM NaCl, pH 5.0), while the spectra of RR14 in buffer with increasing percentages of trifluoroethanol (TFE) exhibited strong minima at 208 and 222 nm and an apparent maximum at 190 nm, indicating the existence of α-helical conformations (Fig. 1A). Additionally, the spectrum for RR14 in 50% TFE showed the maximum amount of α-helical content (41.1%) (Fig. 1A and Table 1). Similarly, increased α-helical propensities of RR14 were observed in the presence of membrane-mimicking solvents (sodium dodecyl sulfate [SDS] and DPC micelles) (Fig. 1B and 1C). The α-helical content of RR14 in the DPC micelles (molar ratio of peptide/DPC of 1:100) was determined to be 42.4% (Table 1). When interacting with SDS micelles (molar ratio of peptide/SDS of 1:100), the α-helical content of RR14 decreased to 35.2% (Table 1). These results demonstrated that RR14 could interact with bacterial membranes through α-helical conformations.

FIG 1.

Far-UV CD spectra of RR14 under TFE, SDS, and DPC buffer conditions. (A) The CD spectra of RR14 in sodium phosphate (control), and 10%, 30%, and 50% TFE are shown in black, salmon pink, red, and brown dot lines, respectively. (B) The secondary structural properties of RR14 in SDS micelles are observed. The CD spectra of RR14 in SDS in molar ratios of 1:100 (light blue), 1:200 (slate), and 1:300 (dark blue) are presented. (C) The CD spectra of RR14 in DPC micelles are displayed in light-green (molar ratio of RR14/DPC = 1:50) and dark-green (molar ratio of RR14/DPC = 1:100) dot lines. MRE, mean residue ellipticity.

TABLE 1.

The α-helical contents of RR14 in different solvent solutions

Solvent	Concn of solvent (%)	Molar ratio of RR14/solvent	α-Helical contents of RR14 (%)
TFE	10		3.3
	20		31.2
	30		38.4
	40		40.9
	50		41.1
SDS		1:100	35.2
		1:200	33.9
		1:300	34.5
DPC		1:50	38.1
		1:100	42.4

Solution structure of RR14 in complex with DPC micelle.

To understand the structural basis for the mode of action of RR14, we further carried out NMR experiments to resolve its solution structure in membrane-mimicking solvents. We acquired several one-dimensional (1-D) and two-dimensional (2-D) NMR spectra of RR14 in both SDS and DPC micelles to identify suitable conditions for structure determination. Low-resolution spectra with poorly dispersed cross-peaks were obtained when RR14 interacted with SDS micelles (Fig. S1 and S2 in the supplemental material), while highly resolved spectra (¹H-¹H 2-D nuclear Overhauser effect spectroscopy [NOESY] and ¹H-¹H 2-D total correlation spectroscopy [TOCSY]) with well-dispersed cross-peaks were acquired when RR14 bound to DPC micelles at pH 5.0 and 318 K (Fig. S3). Consequently, the solution structure of RR14 was determined in DPC micelles. The cross-peaks of ¹H-¹H 2-D TOCSY and ¹H-¹H 2-D NOESY spectra were employed to assign the sequential backbone information. The finished assignments consisted of 230 NOE-derived distance constraints, including 59 intraresidue, 65 sequential, and 106 medium-range distance restraints (Table 2). As well, we derived 24 dihedral angle restraints from 2-D ¹H-¹³C_α heteronuclear single quantum coherence (HSQC) (Fig. 2A) and 2-D ¹H-¹⁵N HSQC (Fig. S4) spectra and combined them with the assigned NOE constraints to calculate the solution structure of DPC micelle-bound RR14 (Table 2). Furthermore, a superimposed ensemble of the 15 lowest-energy structures was obtained and is presented in Fig. 2B. The calculated root mean square deviations (RMSDs) of all heavy atoms and backbone heavy atoms of these lowest-energy coordinates are 0.88 ± 0.24 Å and 0.32 ± 0.11 Å, respectively (Table 2). Moreover, the structural quality of this ensemble was checked by a Ramachandran plot of PROCHECK analysis. The result showed that 94.5% of residues of RR14 are located in the most favored regions; the remaining 5.5% are in the additionally allowed regions. The determined structure revealed that RR14 contains an α-helix (R4–K13) with loop conformations extended at the N and C termini (Fig. 2B). Remarkably, the three-dimensional (3-D) structure of RR14 shown in ribbon diagrams along with the electrostatic surfaces (Fig. 2B and 2C) demonstrated the amphipathic property—the side chains of the hydrophobic residues are on one side, and those of the positively charged residues (R3, R4, K6, R10, R11, and K13) are on the other side of this peptide.

TABLE 2.

NMR structure calculation parameters of RR14^a

Parameter	Value
NOE restraints
Intraresidue NOEs	59
Sequential NOEs $\left[\left(i-j\right)=1\right]$	65
Medium-range NOEs ($\left|i-j\right|$ ≦ 4)	106
Total NOEs	230
Dihedral angle restraints	24
Ramachandran plot summary (%)
Most favored regions	94.5
Additionally allowed regions	5.5
Avg RMSD from the mean structure
Backbone atoms	0.32 ± 0.11 Å
All heavy atoms	0.88 ± 0.24 Å

^aPDB ID 8GVN; BMRB entry 51619.

FIG 2.

NMR spectrum and solution structure of RR14 in DPC micelles (PDB ID 8GVN). (A) The 2-D ¹H-¹³C_α HSQC spectrum of RR14 in DPC micelles. (B) A superposition of the 15 lowest-energy structures of RR14 is shown on the left (the backbones and heavy chains are shown in blue lines). On the right, a ribbon diagram (cyan) representing the averaged DPC-bound solution structure of RR14 is shown; the positively charged residues are shown as balls and sticks in blue. (C) The electrostatic surface of DPC micelle-bound RR14. Positively charged residues are in blue, negatively charged residues in red, and neutral positions in white. N-ter, N terminus; C-ter, C terminus.

Membrane permeabilization of RR14 against synthetic LUVs.

The bactericidal activity of an AMP is mostly associated with its capacity for membrane permeability. Therefore, we evaluated the membrane-disrupting ability of RR14 by using the calcein dye leakage assay. We added RR14 peptide to induce leakage of entrapped calcein from large unilamellar vesicles (LUVs) with distinct surface charge densities. Neutral LUVs were generated with POPC (1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine) lipids to mimic eukaryotic cell membranes. Negatively charged LUVs were made with POPG (1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoglycerol) lipids to mimic bacterial cell membranes. The minimal concentration of peptide that induced 100% leakage of calcein from the LUVs was defined as the LC₁₀₀. The results showed that RR14 induced the leakage of calcein from POPG LUVs with an LC₁₀₀ of 0.5 μM (Fig. 3). However, RR14 was less effective in inducing the leakage from POPC; with a POPC/POPG LUV ratio of 3:1, the LC₁₀₀ values were about 2.0 and 1.0 μM, respectively. The observed weaker potency in breaking the eukaryote-mimetic lipid membranes revealed that RR14 could selectively target and disrupt bacterial membranes.

FIG 3.

The membrane permeabilization activities of RR14 against the POPC, POPG, and POPC/POPG (3:1) LUVs. The membrane permeability of RR14 against POPC liposomes as a function of peptide concentration is shown. The percentages of leakage induced by RR14 are symbolized in green diamonds. The percentages of RR14-induced dye leakage from POPG liposomes as a function of peptide concentration are symbolized in red diamonds. The membrane-disrupting ability of RR14 toward POPC/POPG liposomes in a ratio of 3:1 is estimated and presented in percentages of leakage (blue diamonds). Error bars show standard deviations.

The orientation of RR14 in DPC micelles.

The orientations and localizations of peptides within membrane environments are significantly correlated with their bactericidal mechanisms of action. AMPs can be modified to increase their antimicrobial potencies when the locations and orientations of functionally essential fragments (residues) are known. Furthermore, we employed paramagnetic relaxation enhancement (PRE) experiments to explore the position and orientation of RR14 in DPC micelles. Fatty acids that were spin labeled at the 5th and 12th carbon positions of the acyl chain (5-doxyl-stearic acid [5-DSA] and 12-doxyl-stearic acid [12-DSA]) were used. In the PRE experiment, 5- and 12-DSA perturbed the NH-C_αH signals of residues inserted near the surface and buried inside the micelle, respectively. Subsequently, the acquired NH-C_αH signal (F2 dimension) of each residue of the ¹H-¹H 2-D TOCSY spectra (Fig. S5 to S7) was used to evaluate the relatively decreased intensity in comparison to that of the peptide without spin label effects. The mean remaining amplitudes of RR14 in the presence of 5- and 12-DSA were determined to be 84.2% and 55.8%, individually (Fig. 4A). The remaining amplitudes of residues I5, A7, L9, and R10 were obviously smaller than the average as treated with 5-DSA. Also, residues I5–R10 showed apparently decreased remaining amplitudes compared to the average in the presence of 12-DSA. These results indicate that the I5–R10 segment is most likely buried inside the DPC micelles. In addition, we conducted the PRE experiment in the presence of Mn²⁺ ions to monitor the residues of RR14 that are outsides the DPC micelles and exposed to solvents. It has been reported that Mn²⁺ ions decrease the intensities and broaden the nuclear magnetic resonances of the nearby nucleus through interacting with the phosphate groups of DPC micelles (35). We therefore investigated the PRE effect of Mn²⁺ ions (0.5 mM) on RR14 and observed that the remaining amplitudes of residues R11–A14 showed considerable decreases. This result indicates that the C terminus of RR14 is not buried inside DPC micelles but exposed to the solvent.

FIG 4.

The position of RR14 in DPC micelles. (A) The remaining amplitudes of RR14 in the presence of 5- and 12-DSAs and Mn²⁺ ions. The red, green, and blue lines denote the mean remaining amplitudes of RR14 in the presence of 5- and 12-DSAs and (0.5 mM) Mn²⁺ ions, respectively. (B) The model structure of RR14 in complex with a DPC micelle is deduced from PRE results. The DPC lipids are shown as sticks (light gray), with the sulfur atoms shown as spheres (yellow) and oxygen atoms colored in red. (C) Structural model of RR14 in complex with a DPC micelle (shown in spheres). In both panels B and C, RR14 is presented as a ribbon diagram with the residues colored in magenta (exposed to solvent) and cyan (buried in the DPC micelle).

The membrane-disrupting activity of RR14 against bacteria.

The membrane permeability of RR14 against synthetic LUVs was demonstrated by the calcein dye leakage assay. Furthermore, we evaluated the bioactivity of RR14 in disrupting and permeabilizing bacterial membrane with a propidium iodide (PI) uptake assay. PI, a DNA intercalating dye, can pass through disrupted bacterial membranes (36). PI is able to interact with nucleic acids in the cytoplasm of bacteria when bacterial membranes are permeabilized. The binding to nucleic acids further increases the fluorescent signal of PI molecules (37). Thus, the fluorescence emitted by PI molecules is an indicator of inner membrane permeabilization. We evaluated the damage caused by RR14 (at 10 μM peptide concentration) to bacterial membrane by using the PI uptake assay. The results showed that membrane-disrupting activity of RR14 against Escherichia coli, Acinetobacter baumannii, and S. aureus was apparent compared to that of the control (without peptide addition) (Fig. 5). The hierarchy of relative permeability of RR14 against the tested bacteria is E. coli (relative permeability of 9.0 ± 0.1 [mean ± standard deviation]) > A. baumannii (relative permeability of 6.1 ± 0.1) > S. aureus (relative permeability of 4.2 ± 0.2).

FIG 5.

The membrane permeabilization ability of RR14 against E. coli, A. baumannii, and S. aureus. The membrane-permeabilizing abilities of RR14 against E. coli, A. baumannii, and S. aureus were determined and are presented as relative permeability values. “Control” and “RR14” in each panel indicate that bacteria were treated with PBS and peptide (10 μM), respectively. The relative permeability values determined here for RR14 against E. coli, A. baumannii, and S. aureus are 9.0 ± 0.1, 6.1 ± 0.1, and 4.2 ± 0.2, respectively. Error bars show standard deviations.

DISCUSSION

Nowadays, much effort has been made to develop new classes of therapeutics to combat antibiotic-resistant bacteria (38). Antimicrobial peptides (AMPs) characterized with amphipathic α-helixes in their structures exhibit microbicidal activity against Gram-negative and Gram-positive bacteria and fungi (39, 40). AMPs target bacteria through specific permeabilization of membranes and are of great potential for the development of new antibacterial agents without incurring resistance (41). However, the development of AMPs as feasible therapeutics has been hindered by several problems, including their cost of synthesis, bioavailability, stability, and salt sensitivity (25). In particular, salt sensitivity is a decisive factor for AMPs in their microbicidal mechanism (42). Many AMPs, such as magainins, gramicidins, bactenencins, indolicidins, and defensins, are observed to become inactive in the presence of high salt concentrations (43, 44). Although several researchers have reported investigations of salt-resistant AMPs in inhibiting the growth of bacteria under high-salt conditions, details of structure-function relationships conferring salt sensitivity and explaining the mode of action are not elucidated. Therefore, in this study, we employed biochemical and biophysical methods to explore the mode of action of a novel salt-resistant antimicrobial peptide, RR14, in terms of secondary structure content, membrane-permeabilizing selectivity, solution structure, and position and orientation in complex with DPC micelles.

We employed CD spectroscopy to explore the structural properties associated with the bactericidal activity of RR14. The structural conformation of RR14 in general buffer solution (20 mM phosphate, 100 mM NaCl, pH 5.0) was a random coil, while in 50% TFE, RR14 turned out to be an α-helical structure. This conformational alteration may be caused by the electrostatic attraction from the fluorinated alcohol, which could induce hydrogen bond formation among the amide and carbonyl groups, stabilizing the helical formation. Likewise, in the negatively charged SDS micelles and zwitterionic DPC micelles, RR14 remained in the α-helical conformation. These observations indicate that the conformational transition of RR14 is essential for its ability to interact with and bind to phospholipid membranes. As well, we explored the structural basis underlying the mechanism by which RR14 targets bacterial membranes through resolving the NMR solution structure. To acquire the 1-D and 2-D NMR spectra of RR14, the peptide’s solubility and stability were tested under distinct pH conditions. In an acidic environment (pH 5.0), RR14 was stable and remained soluble as a peptide concentration of more than 2.4 mM, while the solubility of RR14 gradually decreased as the buffer condition become more basic (pH 6.5 to 7.0). Therefore, the solution structure of RR14 in negatively charged SDS or zwitterionic DPC micelles was investigated under the acidic buffer condition (20 mM phosphate, 100 mM NaCl, pH 5.0). The NMR spectra of RR14 in complex with SDS micelles were acquired. However, the 2-D TCOSY spectra obtained (Fig. S2) showed poorly dispersed peaks with low resolution. Generally, the solution structures of AMPs embedded in bicelles, micelles, or nanodiscs are difficult to determine due to instability of the complex and poor quality of NMR spectra (45, 46). It is noteworthy that the cytoplasmic membrane of bacteria consists of both anionic and zwitterionic lipids (47). Upon interacting with cationic AMPs, the anionic lipids are usually observed to cluster, resulting in membrane disruption (48, 49). On the other hand, the charge separation mechanism of AMPs is highly associated with zwitterionic lipids in bacterial membranes (48, 49). Moreover, a number of structures of potent AMPs were resolved by using DPC lipids as a prototypical detergent (50–54). Accordingly, to obtain the 3-D structure of RR14 in a membrane environment, we further conducted NMR experiments on the peptide in complex with zwitterionic DPC micelles. The ¹H-¹H 2-D NOESY and ¹H-¹H 2-D TOCSY spectra obtained for RR14 in DPC micelles (Fig. S3) were highly resolved and better dispersed than those acquired in SDS micelles. Therefore, the solution structure of RR14 in complex with DPC micelles was determined. Structurally, RR14 showed an amphipathic α-helical conformation with residues R4–K13 forming a 2.5-turn α-helix. Thus, residues R4–K13 could be of great importance in contributing to the antimicrobial activity of RR14.

In our dye leakage assay, RR14 was characterized as strongly disrupting POPG liposomes (negatively charged phospholipid vesicles). RR14 induced 100% leakage from POPG, POPC, and POPC/POPG (3:1) with LC₁₀₀ values of 0.5, 2, and 1 μM, respectively, demonstrating the selectivity of RR14 between bacterial and mammalian cell membranes. The resolved α-helical fragment (R4–K13) of RR14 induced by DPC micelles could be significant for the bactericidal ability mediated by membrane permeability. The membrane-permeabilizing ability of AMPs is one of the key mechanisms preventing the development of resistance by pathogens (55, 56). Our PI uptake assays demonstrated that RR14 could target and disrupt the membranes of E. coli, A. baumannii, and S. aureus, leading to apparent observed increases in the fluorescence emitted by PI molecules that bound to the nucleic acids in the cytoplasm of bacteria. In addition, the relative permeabilization abilities of RR14 against E. coli and A. baumannii were observed to be higher than the permeabilization ability against S. aureus (Fig. 5). Of note, E. coli and A. baumannii are classified as Gram-negative bacteria, while S. aureus belongs to the Gram-positive group. The cell walls of Gram-negative bacteria are composed of one thin layer of peptidoglycan, an outer cell membrane, a periplasmic space, and the inner plasma membrane, whereas Gram-positive bacterial cell walls consist of a thick peptidoglycan layer, a periplasmic space, and one lipid plasma membrane. The amphipathic structure of RR14, characterized with charged/polar and hydrophobic moieties (Fig. 2), can integrate into and disrupt the membranes of both Gram-negative and Gram-positive bacteria. However, when targeting Gram-positive bacteria (S. aureus), the access of the peptide to the cytoplasmic membrane could be inhibited by the thick peptidoglycan of the cell wall. This could further reduce the efficiency of RR14 in disrupting the bacterial membrane and result in the lower relative permeability observed in the PI uptake assay.

The amphipathic α-helical conformations of AMPs generally target the bacterial membranes, with their hydrophobic regions binding to the membrane and the polar moiety exposed to the solvents. To further explore the possible binding location and orientation of RR14 in the DPC micelles, PRE experiments were conducted with the additions of Mn²⁺ ion and 5- and 12-DSAs. In the presence of 5- and 12-DSAs, residues I5, A7, L9, and R10 showed reduced peak intensities and broadened resonance signals, resulting in the remaining amplitudes being smaller than the average. The results indicate that the I5–R10 segment is highly likely to be buried in the DPC micelle. This also implies that the buried α-helical segment probably has its hydrophobic residues protruding into the interior of the micelle and the positively charged residues exposed to the inner surface of the micelle. Moreover, the membrane permeability and bactericidal activity of AMPs were reported to correlate with the α-helicity (42, 57). Accordingly, the micelle-induced α-helical segment (R4–K13) is functionally essential for the membrane disruption and bacteriostatic ability of RR14. On the other hand, residues R11–A14 showed considerably reduced remaining amplitudes in the presence of Mn²⁺ ion compared to those in the presence of 5- and 12-DSAs. Therefore, we suggest that the C-terminal residues, R11–A14, are most likely exposed to the solvent. Apart from I5, A7, L9, R10, and R11 to R14, the rest of the residues of RR14 were not apparently affected by Mn²⁺ ion or 5- or 12-DSA; their remaining amplitudes were higher than the average. This could be attributed to the fact that the mobile and constantly changing micelles alter their shapes and positions in solution. Thus, the flexible micelles may hinder the precision of distinguishing the PRE effects on peptides. Eventually, we built a possible model of the RR14-DPC micelle complex (Fig. 4B and C) on the basis of the resolved structure of RR14 and the PRE results.

Mostly, AMPs function to target and disrupt the bacterial membrane, further exerting bactericidal activity, especially small and short peptides characterized with amphipathic helixes interacting with negatively charged microbial membranes by their cationic residues. To understand the possible mechanism of action of RR14, we analyzed its structure and compared it to those of the known short AMPs. We retrieved the structures of peptides CM15, RR12, P3967, and KR12. All these AMPs are short, ranging from 12 to 15 amino acids, and display the amphipathic helical conformation (2- to 3-turn α-helix) (Fig. 6). Consistently, RR14 is also short (14 amino acids) and forms an almost-3-turn α-helix. The AMP CM15 (PDB identifier [ID] 2JMY) is a hybrid peptide derived from melittin(3–9) and cecropin(1–8), P3967 (PDB ID 6RRL) is a peptide identified from the medicinal leech Hirudo medicinalis, KR12 (PDB ID 2NA3) is derived from the human cathelicidin LL37 peptide, and RR12 (PDB ID 6J9P) is a novel salt-resistant antimicrobial peptide. RR12 was first reported by Mohanram and Bhattacharjya (58). The sequence of RR12 was designed to be a 12-residue cationic/hydrophobic peptide with helical amphipathicity (58). It is noteworthy that arginine residues were found to be essential for salt resistance ability in AMPs (34). Thus, arginine was preferentially used for the cationic residues in the designed RR12 sequence (RRLIRLILRLLR). In fact, the bactericidal activity of RR12 functions strongly at high salt concentrations (200 to 300 mM NaCl) (58). The mode of action of RR12 was investigated by our group, and the solution structure of RR12 was determined in the SDS micelle environment (59). Structurally, this peptide presents as an approximately 2.5-turn α-helical conformation (L3–L11) with side chains of arginine residues (R1, R2, R5, R9, and R12) aligned on one side of the overall structure (Fig. 6). Furthermore, by comparing the structures of the above-mentioned short AMPs with that of RR14, we found that the most distinct structural property of these short peptides is the distribution of their cationic residues. In KR12, residues K1, R6, and R12 are on one side of the helix, while R2, K8, and R12 are on the other side. Similar distributions of cationic residues to those of KR12 can be observed in CM15 and P3967. In contrast, in RR12 and RR14, nearly all the positively charged residues (R1, R2, R5, R9, and R12 for peptide RR12 and R3, R4, K6, R10, R11, and K13 for peptide RR14) are oriented on the same side of the amphipathic helix. These observations indicate that the featured arrangements of cationic residues in the structures of RR12 and RR14 most likely contribute to their stable bactericidal activity in high-salt environments. Moreover, arginine residues were observed to contribute to cation-pi interactions associated with improved salt resistance of peptides (43, 44). Thus, the arginine residues of RR14 aligned on the cationic side of the α-helix could render high salt tolerance through hydrogen and ionic interactions with the bacterial membranes.

FIG 6.

Structural comparisons of RR14 with known short AMPs. The solution structure of RR14 (PDB ID 8GVN) is compared with the structures of the following known small and short AMPs: RR12 (PDB ID 6J9P), CM15 (PDB ID 2JMY), P3967 (PDB ID 6RRL), and KR12 (PDB ID 2NA3). The structures of all these peptides are presented as electrostatic surfaces (red, blue, and white denote negative, positive, and neutral charges, respectively). Additionally, the secondary structures of peptides are displayed in ribbon diagrams (yellow), with positively charged residues shown as sticks and labeled.

Conclusions.

In this study, we disclosed the structure-function relationship of a novel salt-resistant AMP, RR14, through biochemical and biophysical techniques. RR14 undergoes a structural transition (random coil to α-helix) to interact with DPC micelles. We resolved the detailed solution structure of RR14 in complex with DPC micelles. RR14 orients toward the interior of the DPC micelle, with its N terminus and the a-helical segment (residues I5–R10) buried inside the micelle, while the C terminus is exposed to solvent. RR14 selectively targets and disrupts negatively charged bacterial membranes, and this is attributed to the central α-helix fragment (I5–R10) which is of great importance for membrane permeabilization and bactericidal activity. The specific and featured arrangement of positively charged residues of RR14 on its amphipathic helical conformation could further render its strong salt resistance ability. Conclusively, our study unveiled the bioactive structure of RR14, explaining its mode of action against MRSA and other microbes. The insights obtained are of great applicability for the development of new antibacterial agents.

MATERIALS AND METHODS

Materials.

The RR14 peptide (WLRRIKAWLRRIKA) was synthesized by Yao-Hong Biotechnology, Inc. (http://www.yh-bio.com.tw/en/index.asp). TFE (2,2,2-trifluoroethanol), MnCl₂, calcein, 5-doxyl stearic acid (5-DSA), and 12-doxyl stearic acid (12-DSA) were purchased from Sigma-Aldrich. D₂O, deuterated methanol (methanol-d4), and dodecylphosphocholine-d38 (DPC-d38) were bought from Cambridge Isotope Laboratories. Sodium dodecyl sulfate (SDS) was purchased from Merk (Darmstadt, Germany). POPC (1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine) and POPG (1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoglycerol) were bought from Avanti Polar Lipids (Alabaster, AL, USA).

Methods.

(i) CD spectroscopy. The Aviv 202 circular dichroism (CD) spectrometer (Aviv, Lakewood, NJ) was employed to perform far-UV CD experiments. The peptide solution stock (2.0 mM) was prepared by dissolving RR14 in 20 mM sodium phosphate, 100 mM NaCl, pH 5.0. The peptide stock was then diluted with solvents TFE, SDS, and DPC separately to the molar ratios of interest for CD experiments. The final concentration of RR14 in the distinct solvents was 60 μM. The tested sample was placed into a 1-mm path length quartz cuvette. Far-UV CD experiments were conducted to record the signals between 190 and 260 nm at 25°C. All samples tested were scanned three times for further averaging and conversion to mean residue ellipticity ([θ]MR; mean molar ellipticity [θ] of individual residues). The BeStSel website (http://bestsel.elte.hu/) (60) was used to estimate the secondary structure contents of all spectra.

(ii) Preparation of LUVs. The generation of large unilamellar vesicles (LUVs) was performed according to a previously described method (61). Phospholipid powders were first solvated in chloroform and thoroughly dried by nitrogen air. Phosphate-buffered saline (PBS) buffer (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, and 1.8 mM KH₂PO₄, pH 7.4) was used to dissolve the dried lipid film, followed by 10 freeze-thaw cycles. Subsequently, a miniextrusion device (Avanti Polar Lipids, Alabaster, AL, USA) was used to extrude the lipid suspensions through 0.4- and 0.1-μm pore-size polycarbonate filters 10 times to generate the LUVs. A similar procedure was used to generate entrapped-calcein LUVs, in which 70 mM calcein was liquefied in 10 mM Tris, pH 7.4, in advance, and the initial solution of entrapped-calcein LUVs was passed through a PD-10 desalting column to remove the unentrapped calcein. Finally, dynamic light scattering on a Zetasizer Nano ZS (Malvern Instruments, Malvern, UK) was employed to check the size of the LUVs generated.

(iii) Calcein leakage assay. The calcein leakage from LUVs induced by the peptide was monitored by measuring the changes in fluorescence intensity. The fluorescence was detected by using an enzyme-linked immunosorbent assay (ELISA) reader (BioTek Synergy H1) with excitation and emission wavelengths of 496 and 515 nm, respectively. The entrapped-calcein LUVs, in a buffer of 20 mM Tris and 100 mM NaCl, pH 7.4, were subjected to peptide-induced leakage at 25°C. As a reference to compare with the results induced by the peptides, 100% leakage of LUVs induced by the addition of 0.1% (vol/vol) Triton X-100 was used. The following equation was used to determine the degree of leakage induced by peptides: % leakage = [(F – F₀)/(F_r – F₀)] × 100, where F and F₀ are the fluorescence intensities obtained with and without peptide addition, respectively, and F_r is the observed fluorescence intensity after Triton X-100 treatment.

(iv) NMR spectroscopy. To prepare the samples for NMR experiments, peptide stock was mixed with DPC-d38 micelles to reach a molar ratio of 1:100 under the buffer conditions of 20 mM sodium phosphate, 100 mM NaCl, and 10% D₂O at pH 5.0. The final concentrations for peptide and DPC-d38 were 1.5 and 150 mM, respectively. The Bruker Avance 800 MHz spectrometer was employed to acquire NMR spectra at 318 K. The natural-abundance 2-D ¹H-¹⁵N HSQC spectrum was collected at 400 points in T₁ and 2,048 points in T₂; the 2-D ¹H-¹³C_α HSQC spectrum was collected at 376 points in T₁ and 1,024 points in T₂ (62). In addition, two distinct mixing times of 150 and 300 ms of ¹H-¹H 2-D NOESY spectra were acquired. As well, the ¹H-¹H 2-D TOCSY spectra were acquired with a mixing time of 80 ms at 320 points in T₁ and 2,048 points in T₂. All the spectra obtained were processed by using NMRPipe (63) and TopSpin 3.1 (Bruker-Spectrospin).

(v) Structure calculation. ¹H-¹H 2-D NOESY experiments with a mixing time of 300 ms were performed to extract the ¹H-¹H distances of RR14 in DPC micelles, and ¹H-¹H 2-D TOCSY (mixing time of 80 ms) and 2-D ¹H–¹H COSY spectra were measured for resonance assignments. The chemical shift assignments and NOE spinlink assignments were carried out manually using CARA (http://cara.nmr.ch/doku.php). According to the peak intensities, the assigned cross-peaks were divided into weak, medium, and strong groups for distance ranges of 1.8 to 2.8, 1.8 to 3.4, and 1.8 to 5.0 Å, respectively. In addition, the chemical shifts of the 2-D ¹H-¹⁵N HSQC and 2-D ¹H-¹³C_α HSQC were employed to predict the backbone dihedral angle constraints by using TALOS (64, 65). Furthermore, the solution structure of RR14 was calculated by using Xplor-NIH via the PONDEROSA server (66–68). Finally, an ensemble of the 15 lowest-energy structural models was validated by using PROCHECK-NMR (69) and MOLMOL (70), with the molecular visualization generated via PyMOL 2.3.4 (http://www.pymol.org).

(vi) PRE experiments. Peptide solutions with the addition of aliquots of 5-doxyl-stearic acid (5-DSA), 12-doxyl-stearic acid (12-DSA), and Mn²⁺ ions were separately subjected to paramagnetic relaxation enhancement (PRE) experiments. MnCl₂·4H₂O powder was liquefied in 20 mM sodium phosphate, 100 mM NaCl, pH 5.0, to prepare the Mn²⁺ ion solution (0.1 M). Subsequently, a mixture of peptide and DPC-d38 micelles (molar ratio of 1:100) was added to the Mn²⁺ ion solution and incubated at room temperature for 15 min before PRE experiments (¹H-¹H 2-D TOCSY). To estimate the PRE effect of DSA, all DSA powders were thoroughly liquefied in deuterated methanol (D₄-MeOH). Later, DPC-d38 solution was mixed with DSA to reach a molar ratio of 1:60 and incubated for 10 min, followed by the addition of peptide. The final concentrations of peptide, DPC, and DSA were 2.4, 197, and 3.28 mM, individually. After equilibration for 15 min, the ¹H-¹H 2-D TOCSY spectra were acquired. T₁ and T₂ of the ¹H-¹H 2-D TOCSY spectra were set to 2,048 and 320, respectively, and acquired at 318 K. The cross-peak intensities of peptide with and without paramagnetic interferences were measured and calculated according to established protocols (71).

(vii) Propidium iodide uptake assay. The propidium iodide (PI) uptake assay was employed to evaluate the membrane permeabilization activity of RR14 against bacteria according to a previous report (72). Briefly, 10 mL wash buffer (10 mM sodium chloride, 10 mM sodium phosphate buffer, pH 7.4) was used to suspend the bacteria (3 mL at 1.25 × 10⁸ CFU/mL). Subsequently, PI was added into the solution with a final concentration of 1 mg/mL. The solution was incubated at room temperature for 5 min, and then peptide was added (10 μM final concentration) and the mixture incubated for 30 min before the measurement of fluorescence. The fluorescence was detected by using an ELISA reader (BioTek Synergy H1) with the excitation and emission wavelengths set to 535 and 617 nm, respectively. The relative permeability was calculated with the following equation: relative permeability = F_{30 min}/F_{0 min}, where the fluorescence intensity (F) values from the indicated times after peptide addition were used.

Data availability.

The solution structure of RR14 determined here can be found in the Protein Data Bank (PDB ID 8GVN). The NMR spectral parameters (chemical shifts and spectral peak lists) of RR14 can be obtained from the Biological Magnetic Resonance Bank (BMRB entry 51619).