A Defensin from the Model Beetle Tribolium castaneum Acts Synergistically with Telavancin and Daptomycin against Multidrug Resistant Staphylococcus aureus

The red flour beetle Tribolium castaneum is a common insect pest and has been established as a model beetle to study insect development and immunity. This study demonstrates that defensin 1 from T. castaneum displays in vitro and in vivo antimicrobial activity against drug resistant Staphylococcus aureus strains. The minimum inhibitory concentration (MIC) of defensin 1 against 11 reference and clinical staphylococcal isolates was between 16–64 μg/ml. The putative mode of action of the defensin peptide is disruption of the bacterial cell membrane. The antibacterial activity of defensin 1 was attenuated by salt concentrations of 1.56 mM and 25 mM for NaCl and CaCl2 respectively. Treatment of defensin 1 with the reducing agent dithiothreitol (DTT) at concentrations 1.56 to 3.13 mM abolished the antimicrobial activity of the peptide. In the presence of subinhibitory concentrations of antibiotics that also target the bacterial cell envelope such as telavancin and daptomycin, the MIC of the peptide was as low as 1 μg/ml. Moreover, when tested against an S. aureus strain that was defective in D-alanylation of the cell wall, the MIC of the peptide was 0.5 μg/ml. Defensin 1 exhibited no toxicity against human erythrocytes even at 400 μg/ml. The in vivo activity of the peptide was validated in a Caenorhabditis elegans-MRSA liquid infection assay. These results suggest that defensin 1 behaves similarly to other cationic AMPs in its mode of action against S. aureus and that the activity of the peptide can be enhanced in combination with other antibiotics with similar modes of action or with compounds that have the ability to decrease D-alanylation of the bacterial cell wall.


Introduction
Antimicrobial peptides (AMPs) are short peptides less than 50 amino acids that are naturally produced in diverse organisms including plants and metazoans in response to pathogen insult [1]. Cationic AMPs are a subcategory of AMPs that are rich in cationic and hydrophobic residues, giving them an overall net positive charge, which in turn enables the AMPs to bind and disrupt the integrity of the negatively charged bacterial cell envelope. In addition to causing membrane disruption, AMPs have also been demonstrated to possess immunomodulatory properties [2]. AMPs have been underexploited as potential therapeutics in the past, but the rising tide of drug resistance in Staphylococcus aureus and other bacterial pathogens, coupled with the shrinking pool of available drugs for treating these infections, has necessitated the need to develop alternative treatment strategies and therapeutic compounds. As a consequence, the use of AMPs is making a resurgence [1].
Staphylococcus aureus has emerged as a leading cause of hospital and community-acquired infections [3]. Although vancomycin is currently used to treat MRSA as an antibiotic of last resort, vancomycin-resistant S. aureus (VRSA) strains have started to emerge, motivating the urgent development of new antibiotics effective against antibiotic-resistant S. aureus. Overall, S. aureus causes approximately 10,800 deaths per year in the United States and approximately 50% of these are due to MRSA. These statistics underscore the urgent need for novel anti-infectives effective against S. aureus.
In the initial part of this study we screened a set of 65 insect-derived AMPs for activity against methicillin resistant S. aureus (MRSA). Only one of the 65 AMPs, defensin 1 from the model beetle Tribolium castaneum (Def1 Genbank accession number: XM_968482) exhibited promising activity against MRSA. The genome of this beetle encodes three defensins, which have been demonstrated to be induced upon septic injury [4]. The aim of this study was to elucidate the mode of action of defensin 1 from Tribolium and to explore whether it displays combinatorial activity with antiobiotics and to test its toxicity against mammalian cells as well as its in vivo efficacy.

Synthesis of insect-derived antimicrobial peptides
Sixty five insect AMPs were synthesized based on publicly available sequence data or sequences generated by transcriptomic analyses of immune-challenged insects, such as T. castaneum, medicinal maggots of the blow fly Lucilia sericata [8], the invasive harlequin ladybird Harmonia axyridis [9], and the greater wax moth Galleria mellonella [10]. These 65 peptides (S1 Table) were produced by solid-phase synthesis (commissioned work by Panatecs GmbH, Tübingen Germany). Confirmation of correct formation of disulfide linkages in synthetic defensin 1 was performed by mass spectrometric analysis of the fragment spectra of disulfide like tryptic peptides (data not shown). The 65 peptides were extracted to 80% purity by reverse-phase chromatography by GenScript (NJ, USA). Following the initial screening, synthetic Tribolium defensin 1 was resynthesized and processed to more than 95% purity, prior to further investigations.

Antimicrobial susceptibility testing
Antimicrobial susceptibility assays were performed in 96 well plates in triplicate using Müller-Hinton broth (Becton Dickinson and Company, NJ, USA) with an assay volume of 100 μl. The 65 AMPs were arrayed in 96 well microplates to a final concentration of 100 μg/ml. The MRSA strain MW2 was grown overnight in TSB and the bacterial inoculum was added to the assay plate at an adjusted initial OD 600 of 0.03. After overnight incubation at 35°C, the absorbance of each well was measured to determine antimicrobial activity. The procedure for determining the minimum inhibitory concentration (MIC) was adapted from an earlier protocol [11]. In brief, two-fold serial dilutions were carried out to get the test compounds in the concentration range 0.0625-64 μg/ml. After overnight incubation at 35°C, the absorbance was measured to determine antimicrobial activity.

Checkerboard assay for combining antimicrobials
The antimicrobial activity of a combination of defensin 1 with either telavancin or daptomycin was determined through a checkerboard assay described in [11]. Briefly, the compounds whose combinations were being tested were arrayed in serial concentrations, vertically for one compound and horizantally for the other compound in the same 96 well microplate. The rest of the procedure involving addition of bacteria and measurement of growth was carried out as described in the previous section for measurement of MIC.

Bacterial cell membrane permeabilization assay
Permeabilization of bacterial membranes was determined by Sytox Green (Life Technologies, CA, USA) uptake by bacterial cells in 96 well plates (CLS3300, Corning, NJ, USA), performed in triplicate. Logarithmically growing MRSA strain MW2 cells were harvested by centrifugation at 4000 rpm for 5 minutes, the pellet was washed twice in PBS and resuspended in PBS to an absorbance of 0.5 at 595 nm. Sytox Green was added to the cells at a final concentration of 5 μM and incubated in the dark for 30 minutes, followed by addition of the cells to compounds serially diluted in PBS. The fluorescence intensities were measured at different time points (excitation = 485 nm, emission = 530 nm).

Human blood hemolysis
The protocol to test the ability of compounds to cause hemolysis of human erythrocytes (Rockland Immunochemicals, PA, USA) was adapted from Rosch et al. [12]. In a 96 well plate, 50 μl of 2% human erythrocytes suspended in PBS was added to 50 μl of compounds serially diluted in PBS and incubated at 37°C for 1 hour. The plate was then centrifuged at 500 G for 5 minutes and 50 μl of the supernatant from each well of the assay plate was transferred to a fresh 96 well plate. Hemolysis was confirmed by both visual observation and measuring absorbance at 540 nm. Experiments were performed in triplicate.

C. elegans-MRSA liquid infection assay
The C. elegans-MRSA liquid infection assay has been described [5]. In brief, the infection assay was performed in standard 384-well assay plates with the test compounds or 1% DMSO as control. S. aureus MW2 was added to the wells at a final OD 600 of 0.04, followed by automated transfer of 15 adult glp-4(bn2);sek-1(km4) worms to each well of the assay plate using a COPAS large particle sorter (Union Biometrica, MA, USA). After 5 days of incubation at 25°C, the worms were stained with the vital dye Sytox Orange after washing away the bacteria. The assay plates were imaged using an Image-Xpress Micro automated microscope (Molecular Devices, CA, USA), capturing both transmitted light and TRITC (535 nm excitation, 610 nm emission) fluorescent images with a 2X objective. The images from the infection assay were processed using the analysis software CellProfiler (http://www.cellprofiler.org/) and the percentage worm survival was calculated for each well of the assay plates. The entire assay was performed in duplicate.

Defensin 1 from T. castaneum inhibits growth of S. aureus
A set of 65 AMPs (S1 Table) synthesized in vitro based on genomic sequences from a variety of insects was tested for inhibition of growth of MRSA strain MW2 in a broth assay at a final AMP concentration of 100 μg/ml. Among the 65 AMPs only defensin 1 from Tribolium castaneum (Fig 1A), inhibited the growth of MW2 cells and the rest of the AMPs behaved similar to the DMSO control and did not restrict bacterial growth as determined by visual and spectrophotometric observation (S1 Table).
Defensin 1 was retested for antimicrobial activity by determining its MIC against staphylococci of diverse genetic backgrounds, including MW2, USA100, USA300, USA400, RN4220, Newman and Staphylococcus epidermis 9142. We also compared the anti-staphylococcal activity of defensin 1 with well known commercially available anti-staphylococcal drugs such as vancomycin, oxacillin and mupirocin. Defensin 1 displayed MICs between 16-64 μg/ml against the tested strains ( Table 1). The antimicrobial activity of defensin 1 was also evaluated against 7 clinical staphylococcal isolates collected by our laboratory (BF 1, 5, 7-11). Among these isolates; BF 1, 5, 7, 8, 10 and 11 demonstrated MICs to oxacillin of >64 μg/ml, whereas defensin 1 displayed an MIC of 32 μg/ml against all tested clinical isolates (Table 1). The antimicrobial activity of defensin 1 is modulated by cations and Dalanylation of bacterial wall teichoic acids (WTAs) The sequence of Tribolium defensin-1 indicates that it belongs to a typical class of insect defensins, which are composed of cyclic peptides possessing cysteine-stabilized α-helix and β-sheet (CSαβ) structural motifs that are crucial for antimicrobial activity. Its six conserved cysteine residues form intramolecular disulfide bridges (Cys1-Cys4, Cys2-Cys5, Cys3-Cys6) that are important for the folding of the peptide. Secondary structure prediction made using the PSIPRED technique (http://toolkit.tuebingen.mpg.de/quick2_d) suggests that defensin 1 forms β-sheets between the residues 39-43 and 78-80. The secondary structure of mature defensin 1 predicted using the heliquest method (http://heliquest.ipmc.cnrs.fr/cgi-bin/ComputParamsV2. py) suggested that the mature peptide formed an α-helix with a net positive charge similar to other cationic AMPs (Fig 1B). Several cationic AMPs, including peptides of both prokaryotic and eukaryotic origin such as indolicidins, gramicidins, bactenecins, and magainins, are known to be sensitive to salt, which lowers their antimicrobial efficacy [13]. The growth of MRSA strain MW2 was examined in a broth assay in the presence of defensin 1 at a concentration of 64 μg/ml along with serial dilutions of NaCl or MgCl 2 . The final concentratiom of the salts ranged from 0.78 to 100 mM. MgCl 2 inhibited the antimicrobial activity of defensin 1 even at low concentrations starting at 1.56 mM. NaCl inhibited the activity of defensin 1 starting from a higher concentration of 25 mM (Fig 2A).
The cell envelope of S. aureus is coated with anionic polysaccharides referred to as wall teichoic acids (WTAs), which play a variety of roles in maintaining bacterial homeostasis [14]. As a form of defense against attack from cationic AMPs, the overall negative charge on the surface of the cell is neutralized by incorporation of D-alanine residues to WTAs, which attenuates the binding of AMPs to the cell envelope. D-alanylation is carried out by the dlt operon consisting of the four genes dltA-D. Earlier studies have shown that abrogation of D-alanylation through disruption of dltA, sensitizes S. aureus to AMPs [14]. Consistent with the conclusion that defensin 1 functions similarly to previously tested cationic AMPs, we found that the MIC of defensin 1 against the wild type Newman strain was 64 μg/ml whereas the MIC against an isogenic Newman ΔdltA mutant [15] was 0.5 μg/ml (Table 1).

Treatment with a reducing agent inhibits the antimicrobial activity of defensin 1
Earlier studies have shown that the antimicrobial properties of AMPs can be inhibited [16] or enhanced [17] by treatment with a reducing agent, such as DTT. In order to study the effect of DTT on the antimicrobial activity of defensin 1, the peptide (final concentration 64 μg/ml) was treated with serial dilutions of DTT adjusted to final concentrations between 0.1 to 100 mM and incubated at room temperature for 1 hour. The effect of DTT alone, native and DTT-treated peptides on the growth of MRSA strain MW2 was examined in a broth assay. DTT by itself displayed a dose dependent growth inhibitory effect on MW2 starting at a concentration of 6.25 mM (Fig 2B). However, at lower concentrations, DTT did not have a direct inhibitory effect on the growth of the bacterium but abolished the antimicrobial activity of defensin 1 at concentrations between 1.56 to 3.13 mM (Fig 2B)." Defensin 1 disrupts the S. aureus cell envelope but does not cause hemolysis of mammalian red blood cells As shown above (Fig 2A), the antibacterial activity of defensin 1 is affected by the overall charge of the cell surface, suggesting that the antibacterial activity of defensin 1 involves physical binding of the AMP to the cell surface. In addition, earlier studies have suggested that several AMPs manifest their antibacterial effect by causing disruptions to the bacterial cell envelope [1]. In order to investigate the antibacterial mode of action of defensin 1, we examined the potential of the peptide to cause disruption of the bacterial cell envelope. MRSA-MW2 cells were exposed to defensin 1 at concentrations between 16-128 μg/ml and bacterial membrane disruption was evaluated by studying uptake of Sytox Green over a period of 30 minutes. Cells treated with defensin 1 displayed a dose dependent uptake of Sytox Green as represented by the increase in cellular fluorescence caused by binding of the dye with the bacterial DNA ( Fig  3A). The control cells treated with DMSO showed no change in cellular fluorescence.
The possibility that defensin 1 might induce damage to eukaryotic cell membranes was tested by examining the ability of the peptide to induce hemolysis. Human red blood cells (RBCs) were treated with serial dilutions of defensin 1 (0-400 μg/ml) or triton X-100 (0.001-1% solution) for 1 hour. Defensin 1 did not cause any RBC hemolysis even at the maximum tested concentration of 400 μg/ml (Fig 3B), which was confirmed by measuring absorbance at 540 nm

Defensin 1 displays synergistic activity with telavancin and daptomycin
The results from the earlier section suggest that defensin 1 manifests its antibacterial activity by causing disruption to the bacterial cell envelope. We evaluated the hypothesis that the peptide might be able to function synergistically with clinically utilized antibacterial agents that also target the bacterial cell envelope. Telavancin and daptomycin are two such antibiotics that are clinically effective against drug resistant S. aureus. Their mode of action involves targetting the bacterial cell membrane [18,19]. A broth microplate assay was used to test for growth inhibition of MRSA strain MW2, in the presence of defensin 1 adjusted to a final concentration between 1-64 μg/ml, in combination with either telavancin or daptomycin adjusted to final concentrations between 0.016-16 μg/ml. The minimum concentration at which the test compounds inhibited bacterial growth both alone and in combinations were determined to calculate the Fractional Inhibitory Concentration (FIC) index according to the equation shown in the methods section ( Table 2). The FIC index was 0.031 and 0.266 for the defensin 1/telavancin and defensin 1/daptomycin combinations, respectively. As the FIC index was less than 0.5 in both cases, this series of experiments suggest that defensin 1 exhibited synergism with both telavancin and daptomycin.

Defensin 1 prolongs worm survival in the C. elegans-MRSA liquid infection assay
The C. elegans-MRSA liquid infection assay is an established model for studying toxicity and in vivo activity of antimicrobial compounds [5]. The in vivo antimicrobial activity of defensin 1 was compared with vancomycin in C. elegans nematodes infected with MRSA strain MW2. As shown on Fig 4, defensin 1 prolonged survival of infected worms at concentrations starting from 12.5 μg/ml and increased the survival rate of infected worms from 22% to 87% as compared to DMSO treated worms. In the case of vancomycin, starting at a concentration of 1.56 μg/ml, the drug prolonged survival of 98% of the infected worms during the duration of the assay.

Discussion
Cationic AMPs are a conserved component of the innate immune system of plants and metazoans [1]. The rising incidence of drug resistance in S. aureus along with the dwindling supply of antibiotics that are effective against these pathogens has led to an interest in AMPs for potential therapeutic applications [1]. The red flour beetle T. castaneum, a worldwide pest of stored food grains, is also an attractive model to study evolution and immunity in insects due to the availability of the full genome sequence and the ability to perform genetic manipulation, such as RNAi gene knockdowns [4]. Bioinformatics, real-time polymerase chain analysis and suppression subtractive hybridization methods have predicted that T. castaneum produces 15 AMPs [20]. To our knowledge, the direct antimicrobial effect of these peptides on pathogenic bacteria have not yet been shown. In this study, we have characterized the antimicrobial properties of the T. castaneum AMP, defensin 1against S. aureus. Defensin 1 causes disruption of the S. aureus cell envelope and the antimicrobial activity of the peptide was inhibited by the presence of monovalent or divalent cations. Reduction of the disulfide bonds in defensin 1 by treatment with DTT, inhibited the antimicrobial activity of the peptide. Defensin 1 displayed synergistic activity with other antibiotics that also target the bacterial cell envelope, such as telavancin and daptomycin. Defensin 1 inhibited growth of S. aureus at a concentration between 16-64 μg/ml (3.3-13.3 μM) with an average MIC of~45 μg/ml (9.3 μM). To our knowledge, the physiological concentration levels of defensin peptides in the haemolymph of T. castaneum have so far not yet been elucidated, but studies with Aedes aegypti adult mosquitoes have shown that the concentration of a defensin peptide drastically increased from negligible levels to as high as 45 μM, 24 hours post-challenge with a bacterial inoculum [21]. Similar observations have been made in humans, where the concentrations of host defense peptides are elevated during pathogen insult. The salivary concentration of human neutrophil peptide (HNP-1) is higher in patients suffering from oral inflammation, than in healthy individuals [22].
Many AMPs manifest their antibacterial activity by causing disruptions to the bacterial cell envelope [1] which imparts certain advantages to AMPs as compared to conventional antibiotics that target metabolic pathways. It would be more difficult for bacteria to develop resistance against a compound having a mode of action that is physical in nature. Moreover, even though defensin 1 displayed a high MIC against MRSA, the peptide was able to cause damage to the bacterial cell envelope and also displayed in vitro synergism with telavancin and daptomycin, two antibiotics that target the bacterial cell membrane. However, the degree of synergism was more inclined towards telavancin, as compared with daptomycin. Possible reasons for this effect could be the difference in the mode of action of the two antibiotics, even though they both target the cell membrane. It has been shown that telavancin causes physical disruption to the bacterial cell membrane apart from affecting cell wall synthesis, whereas daptomycin aggregates at the cell membrane leading to mislocalization of proteins involved in cell division and cell wall synthesis, which in turn leads to cell death [18,19]. Combination therapy with antibiotics that also target the bacterial cell membrane could thus be an attractive option for the therapeutic application of defensin 1.
An additional strategy to improve the antibacterial activity of defensin 1 would be to enhance the binding ability of the peptide with the bacterial cell surface, which in turn would enable the peptide to cause more extensive damage to the bacteria. Bacteria have evolved several strategies to attenuate the effect of AMPs, one of which is neutralization of the overall charge of the bacterial cell surface by D-alanine decoration of the wall teichoic acids, which impedes the binding of AMPs to the cell surface [14]. We found that defensin 1 displayed a 128 fold drop in MIC from 64 to 0.5 μg/ml when tested on a mutant strain that was defective in D-alanylation, possibly due to enhanced ability of the peptide to bind to the bacteria. This finding in turn raises the interesting possibility of enhancing antimicrobial activity of defensin 1 against S. aureus by blocking D-alanylation in the bacteria.
Although the strategies mentioned above might serve to improve the antibacterial activity of defensin 1, an additional factor that needs to be taken into consideration before exploring potential clinical application of defensin 1 is the sensitivity of the peptide to monovalent and divalent cations. In fact, most cationic AMPs are sensitive to high salt concentrations, which act as an impediment to clinical application of cationic AMPs [13]. However, it has been reported that salt resistance in AMPs can be increased by addition of fatty acids, vitamin E, or cholesterol to the termini of AMPs or by replacing tryptophan or histidine residues with the amino acid β-naphthylalanine [13]. It is reasonable to assume that the salt sensitivity of defensin 1 can also be diminished with similar modifications.
To gain further insight into the potential applications of defensin 1, we tested the toxicity of the peptide on human erythrocytes. The absence of any detectable toxicity of defensin 1 against eukaryotic cells also indicates that the peptide should be further investigated in a rodent infection model as a next step. Moreover, the in vivo efficacy of the peptide was tested in the C. elegans-MRSA infection model. Defensin 1 was well tolerated by human erythrocytes and did not cause hemolysis even at the maximum tested concentration of 400 μg/ml. The C. elegans-S. aureus infection model has been widely used to study staphylococcal virulence and pathogenesis since key virulence factors that are important for bacterial pathogenesis in the nematode model are also conserved in humans [23]. Defensin 1 was able to prolong survival of C. elegans infected with MRSA strain MW2 at concentrations starting from 12.5 μg/ml, which is almost 5 fold lower than the in vitro MIC against MW2 (64 μg/ml). The minimum effective concentration (MEC) of the peptide in vivo being lower than the in vitro MIC suggests a possibility that defensin 1 might have immunomodulatory properties. However, the enhanced in vivo activity of defensin 1 could also be due to bioaccumulation of the peptide in the infected worms, or indicate a synergistic or additive effect with AMPs produced by the nematode. A recent study done with AMPs produced in bumblebees has shown that co-occuring AMPs are capable of working synergistically and result in more potent antimicrobial effects at low concentrations [24].
Interestingly, defensin 1 demostrated specific activity against S. aureus and did not display antimicrobial activity against another Gram-positive bacterium such as Enterococcus faecium or the Gram-negative bacteria Klebisella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa and Enterobacter aerogenes. Obvious reasons for the lack of activity against the Gram-negative bacteria could be due to structural and molecular differences within Gram-positive and Gram-negative cell walls. However, it is unclear why defensin 1 had no effect on the Gram-positive E. faecium. Similar variability in the activity of cationic peptides against staphylococcal and enterococcal strains have been reported in an earlier study [25]. Additionally, it is also possible that defensin 1 may have a significant degree of specificity to S. sureus. Synthetic species-specific peptides against Bacillus subtilis, Pseudomonas spp, Escherichia coli, Klebsiella pneumoniae and Salmonella enterica have been successfully generated in the past [26,27].
In conclusion, defensin 1 inhibits growth of S. aureus by targetting the bacterial cell membrane and displays synergism with antibiotics that also target the bacterial cell membrane. Additionally, the peptide did not exhibit any toxic effect on eukaryotic cells at the tested concentrations. These findings validate the antimicrobial potential of insect AMPs and suggest that their clinical potential should be explored further, especially when combined with actimicrobial agents that downregulate D-alanylation of teichoic acids or with agents that disrupt the staphylococcal cell membrane.
Supporting Information S1 Table. Sixty five insect AMPs synthesized based on publicly available sequence data and tested for antimicrobial properties against MRSA strain MW2. (XLSX)