Antimicrobial Peptide Resistance Mechanism Contributes to Staphylococcus aureus Infection

Staphylococcus aureus produces an export pump, Pmt, that provides resistance to human antimicrobial peptides (AMPs) in addition to secreting bacterial toxins. The AMP resistance function significantly promotes disease, providing a proof-of-principle that AMP resistance contributes to bacterial pathogenesis.

Abstract

Antimicrobial peptides (AMPs) constitute an important part of innate host defense. Possibly limiting the therapeutic potential of AMPs is the fact that bacteria have developed AMP resistance mechanisms during their co-evolution with humans. However, there is no direct evidence that AMP resistance per se is important during an infection. Here we show that the Staphylococcus aureus Pmt ABC transporter defends the bacteria from killing by important human AMPs and elimination by human neutrophils. By showing that Pmt contributes to virulence during skin infection in an AMP-dependent manner, we provide evidence that AMP resistance plays a key role in bacterial infection.

Antimicrobial peptides (AMPs) form an important part of the innate immune system of humans and many other organisms [1]. They significantly contribute to the immune defenses on mucosal and epithelial surfaces. Furthermore, they constitute a key part of the nonoxygen-dependent killing of bacteria inside the phagosomes of professional phagocytes such as neutrophils. In our age of increasing and multiple antibiotic resistance, AMPs have frequently been proposed as promising alternatives to traditional antibiotics [2]. However, to assess the potential of AMPs as antibacterial drugs, one has to consider the bacterial resistance mechanisms that have arisen during the long evolutionary interaction with higher eukaryotes [3].

Most AMP resistance mechanisms work by adjusting basic cellular features [4]. This is also true in the case of the major human pathogen Staphylococcus aureus, which combats AMPs predominantly using nonspecific secreted proteases and alteration of the cell surface charge [5–7]. The latter is accomplished by introducing positively charged phospholipids and modifying teichoic acids with d-alanine, resulting in a positive net charge that repels the commonly cationic AMPs [6, 7]. Because these general physico-chemical alterations of the cell surface have multiple consequences in addition to increasing the repulsion of AMPs, it is difficult to attribute the phenotypes conferred by the respective genes specifically to AMP resistance. This is especially true for investigations of in vivo significance. As a consequence, it remains unclear whether AMP resistance mechanisms are beneficial for bacteria during human infection.

Efflux pumps are an efficient way for bacteria to dispose of harmful substances [8]. However, in staphylococci and other important Gram-positive pathogens, most known peptide efflux pumps are specific for bacteriocins or peptide antibiotics [4]. Gram-positive efflux pumps that contribute to bacterial pathogenesis by broadly protecting them from important human AMPs have not been described. We recently discovered the secretion system of the phenol-soluble modulin (PSM) family of staphylococcal peptide toxins [9]. This system, a 4-component ABC transporter called Pmt, secretes all members of the PSM family. Divergent in amino acid sequence and length, PSMs only have pronounced α-helicity and amphipathy in common, features that underlie their pore-forming activity [10]. Prompted by the general similarity of many AMPs with PSMs in structure and mode of action, we evaluated whether Pmt is a general export system for membrane-damaging peptides, secreting host-derived AMPs in addition to the bacterial PSM cytotoxins.

METHODS

Study Approval

The Animal Care and Use Community at National Institute of Allergy and Infectious Diseases (NIAID), National Institutes of Health, reviewed and approved the used animal protocol (LB-1E), according to the animal welfare act of the United States (7 U.S.C. 2131 et. seq.). All mouse experiments were performed at the animal care facilities of the NIAID in accordance with approved guidelines. Euthanasia was performed with CO₂ at the end of all animal studies. Human neutrophils were isolated from venous blood of healthy volunteers in accordance with protocols approved by the Institutional Review Board for Human Subjects, NIAID. Informed written consent was obtained from all volunteers.

Chemicals and Reagents

LL-37 was purchased from Invivogen, and human beta defensin 3 ([hBD3] DEFB103A) was purchased from Creative BioMart. The hBD3 was resuspended in water, LL-37 in 0.01% acetic acid. Antimicrobial peptides were stored at −20°C until needed.

Bacterial Strains and Culture Conditions

The construction of the completely PSM-deficient isogenic mutant (with deletions of the psmα and psmβ operons and a start codon mutation of the hld gene abolishing production of δ-toxin) in the S aureus USA300 LAC background strain and the Walker A and Walker B mutation-harboring pmt expression plasmids have been described previously [9, 11]. The genomic pmt and pmt/Walker mutants were produced in this study by polymerase chain reaction (PCR) amplifying the pmtRABCD operon plus 1-kilobase pair flanking sequences and ligating into pIMAY, introducing pmtABCD deletion or Walker site mutations, respectively, by inverse PCR (see Table 1 for oligonucleotides used). After transformation into the psm deletion strain, the allelic replacement procedure was performed. The fidelity of all constructs was confirmed by deoxyribonucleic acid sequencing.

Table 1.

Oligonucleotides Used in this Study

Name	Sequence
pmt-1kb Forward	TGGTACCAGACTTGGGGTTG
pmt-1kb Reverse	ACGGTAAGATCAGGGTGAGC
∆pmt Reverse	GATGATTCCTCCTCATAAATGAAC
∆pmt Forward	ATGATAAAAATAATTTTGAGGTTGGGAA
pmtA K41M Forward	ATGACCACAATAATAAGGTTAATTATGGATTT
pmtA K41M Reverse	ACCAGCGCCATTTCTACCA
pmtC K39M Forward	ATGACAACCGTTATGAAAGTAATGAATG
pmtC K39M Reverse	ACCAACACCGTTTTTTCCTA

For in vitro and animal experiments, overnight cultures of bacteria were inoculated at 1:100 dilution into 50 mL tryptic soy broth (TSB) in 125-mL baffled flasks and grown at 37°C shaking at 180 rpm, to mid-exponential growth phase (~2 to 3 hours). Bacteria were then harvested, washed, and diluted with in the appropriate buffer, as described below before experiments. Chloramphenicol was added to plasmid-harboring strains at 10 µg/mL.

Antimicrobial Peptides Assays

For survival of bacteria with AMPs (killing assays), precultures grown overnight in TSB were inoculated into 50 mL of filter-sterilized specialized media (20% [w/v] TSB, 10% [v/v] fetal bovine serum, 1 mM NaH₂PO₄, 150 mM NaCl, 50 mM NaHCO₃; pH 7.4) to an optical density at 600 nm of 0.1 and grown at 180 rpm at 37°C for 2 hours. The bacteria were harvested and adjusted to 2 × 10⁸ colony-forming units (CFUs)/mL. To each well of a 96-well round-bottomed plate, 50 µL of bacteria and 50 µL of AMP were added (LL-37, 180 μg/mL; hBD3, 75 μg/mL). The plate was incubated at 37°C shaking at 180 rpm for 5 minutes and then transferred into a 37°C incubator for 3 hours. Serial dilutions of the incubations were plated on TSA plates. After incubation at 37°C overnight, CFUs were counted the next day.

Neutrophil Experiments

Human neutrophils were isolated from venous blood of healthy donors. The heparinized blood was incubated with an equal volume of 3% (v/v) dextran (Sigma) for 20minutes to separate the erythrocytes. The clear supernatant above the erythrocytes was aspirated, and leukocytes were collected by centrifugation. The cell pellet was resuspended in 0.9% NaCl and layered with Ficoll Hypaque Plus (GE Healthcare). After centrifugation, the supernatant was discarded, and the remaining cell pellet was subjected to a brief hypotonic shock with pyrogen-free water. The cells were washed and resuspended in Roswell Park Memorial Institute (RPMI) medium without phenol red supplemented with 10 mM (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid [RPMI-H]) to the desired concentration.

For determination of bacterial intracellular survival in neutrophils, approximately 3–5 × 10⁷ CFUs/mL for a final multiplicity of infection (MOI) of 10:1, resuspended in RPMI-H, were opsonized with 10% (v/v) prewarmed human serum, incubated for 20 minutes at 37°C under slow rotation, and washed twice with RPMI-H. Then, 3–5 × 10⁶ CFUs were incubated with 3–5 × 10⁵ neutrophils in sterile 2-mL Eppendorf tubes for 20 minutes at 37°C under slow rotation to allow phagocytosis to occur. After 20 minutes, gentamicin (400 µg/mL; Invitrogen) was added to kill extracellular bacteria. To 20 incubations per strain gentamicin was added after the phagocytosis step. Five of those were processed immediately to assess numbers of phagocytosed bacteria (0-hour control). The other 15 incubations, which were allowed to incubate for an additional 60 minutes, were assessed for intracellular CFUs. A 40-µL aliquot was taken from each neutrophil reaction and diluted into 500 μL RPMI-H. The samples were centrifuged, the supernatants were removed, and sterile water was added to the bacterial/cell pellet to promote host cell lysis and release intracellular bacteria. The lysate was immediately plated onto TSA, and CFUs were enumerated after overnight incubation at 37°C.

Mouse neutrophils were isolated and purified from the bone marrow of mouse tibias and femurs as described [12]. The isolated neutrophils were resuspended in RPMI-H at a concentration of 1 × 10⁶ cells/mL, and bacteria were prepared as described above also in RPMI-H to a concentration of ~1.5 × 10⁸ CFUs/mL. A volume of 50 µL of bacteria was added to 150 µL of neutrophils for a final MOI of 50:1 in 96-well round bottom plates. The plates were centrifuged at 250 ×g for 7 minutes at rom temperature and then incubated for 30 minutes at 37°C. After 30 minutes, gentamicin (400 μg/mL; Invitrogen) was added for 60 minutes to kill extracellular bacteria. The cells were washed twice with phosphate-buffered saline (PBS) and finally resuspended with 180 µL PBS and 20 µL of a 1% Triton X-100 solution to lyse the neutrophils and release intracellular bacteria. The lysate was immediately plated onto TSA, and CFUs were enumerated after overnight incubation at 37°C.

Mouse Skin Infection Model

Age-matched female (between 6 and 10 weeks old) C57BL/6J mice and mice deficient in the production of cathelicidin (B6.129X1-Camp<tm1Rlg>/J) were purchased from Jackson Laboratories. The dorsa were shaved, and fine hair was removed with the application of depilatory cream. A total of ~1×10⁸ CFUs of bacteria in 50µL PBS were injected subcutaneously into the left flank. An electronic caliper was used to measure the length (L) and width (W) of the abscess or lesion caused by the bacterial infection each day postinfection. The total size of the abscess was calculated using the formula L×W. To determine CFUs in abscesses 24 hours after infection, the total abscess on each mouse was surgically removed and cut into small pieces with sterile surgical scissors and split into 4 separate tubes containing 500 µg of borosilicate glass beads and 500 µL sterile PBS. The skin pieces were homogenized in a FastPrep-96 (MP Biomedicals) homogenizer, for 5 × 1 minute at 1800 rpm. The homogenates were then plated onto TSA plates and incubated at 37°C, and CFUs were enumerated the following day. For this study, investigators were blinded to the nature of the bacterial strain applied.

Peptide Structures

Structural data were obtained from National Center for Biotechnology Information and loaded into Protein 3D (Lasergene) for visualization.

Statistics

Statistical analysis was performed using Graph Pad Prism version 6.0 with one-way or two-way analysis of variances. All error bars depict the standard deviation (±SD). Lines depict the mean.

RESULTS

The Pmt Secretion System Is an Antimicrobial Peptides Efflux Pump

To investigate a putative AMP export function of Pmt, we produced isogenic mutants of the PSM-deficient derivative (∆αβhld) of the clinically important, methicillin-resistant S aureus strain LAC (pulsed-field type USA300) [11]. The PSM-deficient background was used because (i) in the presence of PSMs, Pmt cannot be deleted, as PSMs accumulate in the intracellular space and kill the bacteria [9], and (ii) the absence of PSMs allows defining the specific role of the presumed AMP export function—compared with the PSM export function—of Pmt. In addition, we constructed a mutant in which the conserved ATP-binding regions (Walker A sites) [13, 14] were mutated in the pmtA and pmtC genes, to demonstrate that the detected phenotypes were due to the catalytic (transport) function of Pmt. We found that Pmt contributed to resistance toward 2 important human AMPs, hBD3 and LL-37 (Figure 1A and B). The hBD3 is produced by neutrophils and keratinocytes and, in contrast to many other β-defensins, maintains anti-staphylococcal activity at physiological salt concentrations; it is deemed crucial for controlling staphylococcal colonization of epithelial surfaces [15]. The cathelicidin LL-37 is a proteolytic product of the hCAP-18 precursor protein [16] and is produced by several cell types including neutrophils [17]. These AMP substrates of Pmt are membrane-active, pore-forming peptides, but they have highly divergent structures; hBD3 is predominantly composed of beta-strands [18], whereas LL-37 forms an amphipathic α-helix [19]. Solution structures of LL-37 and hBD3 in comparison to exemplary structures of PSMα and PSMβ peptides are shown in Figure 2. Only PSMα and LL-37 share clearly recognizable similarity, because they both are composed essentially of 1 α-helix. PSMβ and hBD3 are more complicated in structure, and no clear similarities between them or with the other peptides can be detected. It needs to be stressed, however, that the shown structures are solution structures [18, 20], except for that of LL-37, which was determined in lipid micelles [21]. The conformation adopted by the PSM and hBD3 peptides after insertion into the membrane may be different. In any case, these considerable structural differences suggest that membrane insertion rather than a specific structure makes AMPs substrates of Pmt. More importantly, in all experiments, the Walker site mutant behaved similar to the complete pmt deletion mutant, demonstrating that the detected phenotypes are due to genuine transport. Furthermore, the phenotypes observed in the pmt mutant could be restored by genetic complementation, but not when either Walker site was mutated on the plasmid-encoded pmt, confirming the involvement of Pmt and genuine transport (Figure 1C and D). Of note, it is not surprising that Pmt accepts membrane-inserting substances that originate from either the extra- or intracellular milieu, ie, AMPs and PSM cytotoxins, respectively, because it is a commonly accepted view that efflux pumps of membrane-inserting substances can bind their substrates within the membrane in what has been called a “vacuum cleaner” mechanism [22] (Figure 3).

Figure 1.

Export by Pmt provides resistance to human antimicrobial peptides (AMPs). (A and B) Killing assays were performed in triplicate in phenol-soluble modulin (PSM)-deficient (∆αβhld), PSM/Pmt-deficient (∆αβhld∆pmt), and PSM/Pmt Walker site-mutated (∆αβhld/pmtWalker) isogenic strains in the LAC (USA300) background. *, P < .05; **, P < .01; ***, P < .001 (two-way analysis of variance [ANOVA] with Tukey’s posttest). Error bars show the mean ± standard deviation (SD). (C and D) The LAC ∆αβhld∆pmt strain was complemented with plasmids harboring wild-type or ATP-binding motif (Walker A, Walker B) mutated pmt genes. Killing assays, LL-37, 180 µg/mL; human beta defensin 3 (hBD3), 75 µg/mL. *, P < .05; ***, P < .001; ****, P < .0001 (n = 5/group; one-way ANOVA with Dunnett’s posttest versus values in the pmt mutant). Error bars show the mean ± SD. Abbreviations: CFU, colony-forming units.

Figure 2.

Structures of antimicrobial peptides (AMPs) used in this study. Structures of phenol-soluble modulin (PSM)α3 (Protein Data Bank indentification [PDB ID] 5KGY) and PSMβ2 (PDB ID 5KGZ) as examples of α- and β-type PSMs, respectively, and of LL-37 (PDB ID 2K6O) and human beta defensin 3 ([hBD3] PDB ID 1KJ6). N termini are placed at the (bottom) left. Structures were obtained from the structural data bank at the National Center for Biotechnology Information. α-helices and β-sheets are shown as ribbons. Red and blue colors represent positive and negative charges, and yellow color represents sulfur atoms. Note that these are solution structures (except for LL-37, obtained in lipid micelles), whereas the confirmation during membrane insertion may be considerably different.

Figure 3.

Model of Pmt function. Pmt secretes both the phenol-soluble modulin (PSM) cytotoxins the bacteria produce and host-derived antimicrobial peptides (AMPs). Although PSMs originate from the intracellular and AMPs originate from the extracellular environment, both are membrane-inserting substances that current models predict bind to a hypothetical membrane-located binding site that is common for the efflux pump family that exports such substrates.

Pmt Provides Resistance From Elimination by Antimicrobial Peptides-Producing Human Neutrophils

Being produced in particular by phagocytes, AMPs significantly contribute to our immune defenses against invasive [23]. Neutrophils in particular are the main cells that eliminate invading S aureus after phagocytosis [24]. Pmt had a significant AMP-dependent contribution to bacterial resistance toward killing inside neutrophils (Figure 4). To confirm that this difference was due to diminished killing rather than uptake, we performed a control experiment showing no differences in bacterial numbers at the 0-hour control time point (Figure 4).

Figure 4.

Pmt promotes survival in human neutrophils. Resistance to neutrophil killing was determined 1 hour after completed phagocytosis (0-hour control) and removal of remaining extracellular bacteria (0 hour, n = 5; 1 hour, n = 15). *, P < .05; **, P < .01 (one-way analysis of variance with Dunnett’s posttest versus ∆αβhld). Error bars show the mean ± standard deviation. Abbreviations: CFU, colony-forming units; N.S., not significant.

The Antimicrobial Peptides Resistance Function of Pmt Contributes to Staphylococcus aureus Skin Infection

To directly assess whether the AMP-dependent contribution of Pmt to cellular immune evasion mechanisms translates to an impact on infection, we used a murine skin infection model. Pmt had a significant, AMP-dependent contribution to S aureus virulence during skin infection (Figure 5). These results demonstrated that there is a significant contribution of the PSM-independent AMP export function to S aureus infection, which adds to the established role the Pmt-exported PSMs play in S aureus-mediated disease [10].

Figure 5.

Pmt promotes bacterial survival during skin infection in an antimicrobial peptide-dependent fashion. C57BL/6J mice or cathelicidin-related antimicrobial peptide (CRAMP)-deficient mice were given injections subcutaneously with 10⁷ bacteria. Abscess sizes were measured daily, and colony-forming units (CFU) were determined at day 1 (24 hours) after infection (p.i.) in n = 10 mice/group. *, P < .05; **, P < .01; ****, P < .0001 (two-way analysis of variance with Tukey’s posttest). Error bars show the mean ± standard deviation. Note that no comparisons in the experiment with CRAMP-deficient mice were significant (P ≥ .05). Abbreviations: N.S., not significant; WT, wild type.

DISCUSSION

Our results obtained in a PSM-deficient background strongly suggested that the AMP export function of Pmt underlies the observed phenotypes. To clearly demonstrate whether an AMP resistance mechanism matters for bacterial infection, the most difficult challenge is to link it to AMPs in vivo, a problem that is mostly due to the multitude of AMPs produced in humans and mice. In the past, streptococcal skin infection in mice that were deficient in the cathelicidin-related antimicrobial peptide (CRAMP) equivalent of LL-37 was used to demonstrate that AMPs confer protection from bacterial infection [25]. These results indicated that there is a pronounced importance of CRAMP/LL-37 for the elimination of bacteria in the skin. The significant, non-PSM-dependent impact of Pmt on abscess formation in wild-type mice that we had observed was absent in CRAMP-deficient mice, as determined by abscess sizes measured daily for 7 days (Figure 5A) and CFU at the day 1 (24 hours) maximum of abscess formation (Figure 5B). It is notable that the abscess sizes in the CRAMP-deficient mice were in the same range as those in the wild-type mice infected with the wild-type (∆αβhld) bacteria, independently of Pmt presence or functionality, as one would expect. These results confirm that the non-PSM-dependent in vivo phenotype of Pmt is indeed due to AMP resistance.

To confirm that the absence of significant differences in CRAMP-deficient mice is due to the abolished contribution of CRAMP/LL-37-mediated killing of bacteria inside neutrophils, as our experiment with human neutrophils indicated (Figure 4), we measured survival of bacteria inside neutrophils obtained from wild-type and CRAMP-deficient mice. Significant differences between the Pmt-deficient as well as Walker site-mutated strains and the wild-type bacteria were only observed with neutrophils from wild-type but not CRAMP-deficient mice (Figure 6).

Figure 6.

Pmt promotes survival in mouse neutrophils in an antimicrobial peptide-dependent fashion. Resistance to elimination by mouse wild-type (WT) versus cathelicidin-related antimicrobial peptide (CRAMP)-deficient neutrophils was determined after 1 hour of incubation after synchronized phagocytosis. n = 5/group. *, P < .05 (one-way analysis of variance with Dunnett’s posttest versus ∆αβhld). Error bars show the mean ± standard deviation. Abbreviations: CFU, colony-forming units; N.S., not significant.

CONCLUSIONS

In conclusion, in this study, we describe an efflux pump in the major human pathogen S aureus that provides resistance to important human AMPs. Our findings emphasize the value of Pmt as a target for the development of antistaphylococcal therapeutics, because Pmt combines 2 key immune evasion mechanisms, AMP resistance and cytotoxin secretion. More importantly, our results provide a proof-of-principle for the importance of AMP-resistant mechanisms during bacterial infection.

Notes

Acknowledgments. We thank Nana Ama Amissah for technical assistance.

Author contributions. G. Y. C. C., E. L. F., J. W. M., J. C., J. W. M. C., S. W. D., and M. O. performed research. G. Y. C. C., S. W. D., and M. O. designed experiments. G. Y. C. C., J. W. M., and M. O. analyzed data. G. Y. C. C., S. W. D., and M. O. wrote the paper.

Financial support. This study was funded by the Intramural Research Program of the National Institute of Allergy and Infectious Diseases (project number ZIA AI000904; to M. O.) and the Postdoctoral Research Associate Program of the National Institute of General Medical Sciences (1FI2GM11999101; to S. W. D.), US National Institutes of Health.

Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.