Selected Antimicrobial Peptides Inhibit In Vitro Growth of Campylobacter spp.

: Campylobacter is a major cause of acute human diarrheal illness. Broiler chickens constitute a primary reservoir for C. jejuni leading to human infection. Consequently, there is a need for developing novel intervention methods. Antimicrobial peptides (AMP) are small proteins which have evolved in most lifeforms to provide defense against microbial infections. To date, over 3000 AMP have been discovered; however, few of them have been analyzed speciﬁcally for ability to kill campylobacters. We selected and evaluated a set of 11 unique chemically synthesized AMP for ability to inhibit growth of C. jejuni. Six of the AMP we tested produced zones of inhibition on lawns of C. jejuni . These AMP included: NRC-13, RL-37, Temporin L, Cecropin–Magainin, Dermaseptin, and C 12 K-2 β 12 . In addition, MIC were determined for Cecropin–Magainin, RL-37 and C 12 K-2 β 12 against 15 isolates of Campylobacter representing the three most common pathogenic strains. MIC for campylobacters were approximately 3.1 µ g/mL for AMP RL-37 and C 12 K-2 β 12 . MIC were slightly higher for the Cecropin–Magainin AMP in the range of 12.5 to 100 µ g/mL. These AMP are attractive subjects for future study and potential in vivo delivery to poultry to reduce Campylobacter spp. populations.


Introduction
Antimicrobial peptides (AMP) are small proteins which have been found in almost every class of living organism where they have evolved as a host-defense mechanism against invading microorganisms [1]. For this reason, they are sometimes referred to as host-defense peptides and are considered key players in innate immunity [2]. AMP represent a unique and diverse group of molecules which can be divided into subgroups on the basis of their amino acid composition and structure [3]. They are generally short, 10-40 amino acids in length, cationic, amphipathic, and can be arbitrarily categorized according to their secondary structure as alpha-helices, beta-sheet, extended helices, and loops [4]. A growing number of AMP are being discovered and synthetic new ones are being designed thanks to tremendous research efforts in response to the dramatic and continued evolution of antibacterial-resistant strains of bacteria and the resulting international crisis in health care [1]. For example, a 2004 database contained 523 known peptides, whereas a 2013 database held 3904 natural and 1643 synthetic ones [5,6]. More recently, a 2021 database contains 22,259 entries which include 5891 general AMP, 16,110 patent AMP, and 77 AMP currently in preclinical or clinical stages of drug development [7]. While it is reasonable to expect that some of the AMP that have been discovered or designed de novo to kill a variety of Gram (−) bacteria might also inhibit G (−) Campylobacter spp., there have been very few reports in the literature describing effects of AMP on these specific pathogens, perhaps due to the relative difficulty of culturing campylobacters under microaerobic conditions necessary for its growth.

Materials and Methods
We selected and evaluated a set of 11 unique chemically synthesized and commercially available AMP (Tables 1 and 2) for ability to inhibit growth of C. jejuni isolates 11,168 and 81-176 and seven additional pathogenic bacterial isolates including two Clostridium perfringens isolates, two Salmonella isolates, two Listeria monocytogenes isolates, and one E. coli O157:H7 in a spot-on-lawn assay. Peptides were synthesized using standard solid-phase (Fmoc) chemistry with a peptide synthesizer (CPC Scientific Inc., Sunnyvale, CA 94089, USA, C 12 K-2β 12 ; AnaSpec, Fremont, CA 94555, all other AMP).    [50] (Sigma-A6013) using a swab dipped in the suspension once for the first half the plate, and again for the second half. The inoculated plates were air-dried for approximately 10 min and were then spotted with 10 µL of each antimicrobial peptide or sterile water (control). All AMP solutions were prepared as 1 mg/1 mL; therefore, a 10 µL spot equates to 10 µg. Plates with AMP spots were allowed to dry for approximately 5 min, then were incubated at 37 • C under microaerobic conditions. AMP efficacy was determined as visible zones of inhibition of growth at 24 h.
Six of the AMP we tested produced zones of inhibition on lawns of C. jejuni. Three AMP were chosen for further investigation on the basis of anti-Campylobacter activity, water solubility, and reported reduced cytotoxicity to mammalian cells. Cecropin-Magainin, RL-37, and C 12 K-2β 12 were tested for ability to produce zones of inhibition in spot-on-lawn assays against 24 different bacteria including C. jejuni, C. coli, and C. lari isolates as well as two strains of Salmonella, Clostridium perfringens, Listeria monocytogenes, Lactobacillus, and E. coli O157:H7. The spot-on-lawn assay was conducted as described earlier with the addition of Brain Heart Infusion agar with 1.5% EEO agarose for salmonellas, E. coli, clostridia, and Listeria. Lactobacillus MRS agar with 1.5% EEO agarose was utilized for lactobacilli. Campylobacter and lactobacilli were incubated in a microaerobic atmosphere (10% CO 2 , 5% O 2 , 85% N 2 ) at 37 • C, and ambient air cultivation at 37 • C was utilized for Salmonella, E. coli, and Listeria. An anaerobic environment (5% CO 2 , 5% H 2 , 90% N 2 ) at 37 • C was used for incubation of C. perfringens.
In addition, a modification of the CLSI M26A [51] and Wu and Hancock [52] assays were utilized to determine minimum inhibitory concentrations (MIC) for the AMP Cecropin-Magainin, RL-37, and the oligo-acyl-lysyl (OAK) C 12 K-2β 12 against 15 isolates of campylobacter in microtiter plates. Briefly, target campylobacter cultures were grown in Brucella broth for 18 h in 25 cm 2 vented tissue culture flasks in a microaerobic gas environment (5% O 2 , 10% CO 2 , 85% N 2 ) at 37 • C. Visual adjustment of suspensions equivalent to a 0.5 McFarland standard for 1.5 × 10 8 cfu/mL baseline were made using fresh Brucella broth. The standardized suspensions were further diluted in fresh Brucella broth 1:1000 for an inoculum of 1.5 × 10 5 . Enumeration of inoculum levels was assessed on BAB. Cecropin-Magainin, RL-37, and C 12 K-2β 12 antimicrobial peptides were serially diluted in polypropylene strip tubes using 0.2% bovine serum albumin and 0.1% acetic acid diluent. An amount of 10 µL of each peptide dilution was added to 90 µL of bacterial inoculum culture in Greiner 96-well V-bottom polystyrene tissue culture-treated microtiter plates in triplicate for targets. One well was inoculated for each target with 10 µL 0.2% bovine serum albumin and 0.1% acetic acid diluent with 90 µL of bacterial inoculum as positive control. One well was inoculated with 10 µL 0.2% bovine serum albumin and 0.1% acetic acid diluent with 90 µL of Brucella broth as negative control. Plates were placed in gallon-sized zip-lock freezer bags (SC Johnson, Racine WI 53403, USA), flush-filled with the microaerobic gas mixture, and incubated for 40 h with gas replacement at 24 h. MIC was defined as the minimum concentration that prevented growth based on visual observation using mirrored white light.

Results
A set of 11 unique AMP were chemically synthesized and evaluated for ability to lyse C. jejuni in a spot-on-lawn assay ( Figure 1). Table 3 shows the initial screening results for the 11 AMP against two isolates of C. jejuni as well as additional human foodborne pathogens including isolates of Clostridium perfringens, Salmonella, Listeria, and enterotoxigenic E. coli. Six of the selected AMP produced obvious zones of inhibition against growth of C. jejuni isolates. These antimicrobial peptides included: NRC-13 Pleurocidin, RL-37, Temporin L, Cecropin-Magainin, Dermaseptin, and C 12 K-2β 12 . Interestingly, these same six AMP were also able to lyse most of the other pathogenic bacteria tested. amount of 10 µL of each peptide dilution was added to 90 µL of bacterial inoculum culture in Greiner 96-well V-bottom polystyrene tissue culture-treated microtiter plates in triplicate for targets. One well was inoculated for each target with 10 µL 0.2% bovine serum albumin and 0.1% acetic acid diluent with 90 µL of bacterial inoculum as positive control. One well was inoculated with 10 µL 0.2% bovine serum albumin and 0.1% acetic acid diluent with 90 µL of Brucella broth as negative control. Plates were placed in gallon-sized zip-lock freezer bags (SC Johnson, Racine WI 53403, USA), flush-filled with the microaerobic gas mixture, and incubated for 40 h with gas replacement at 24 h. MIC was defined as the minimum concentration that prevented growth based on visual observation using mirrored white light.

Results
A set of 11 unique AMP were chemically synthesized and evaluated for ability to lyse C. jejuni in a spot-on-lawn assay (Figure 1). Table 3 shows the initial screening results for the 11 AMP against two isolates of C. jejuni as well as additional human foodborne pathogens including isolates of Clostridium perfringens, Salmonella, Listeria, and enterotoxigenic E. coli. Six of the selected AMP produced obvious zones of inhibition against growth of C. jejuni isolates. These antimicrobial peptides included: NRC-13 Pleurocidin, RL-37, Temporin L, Cecropin-Magainin, Dermaseptin, and C12K-2β12. Interestingly, these same six AMP were also able to lyse most of the other pathogenic bacteria tested.
Cecropin-Magainin, RL-37, and C 12 K-2β 12 were chosen for further investigation on the basis of anti-Campylobacter activity, water solubility, and limited cytotoxicity to mammalian cells and were tested for ability to produce zones of inhibition in spot-onlawn assays against 24 different bacteria including 15 unique C. jejuni, C. coli, and C. lari isolates as well as two strains of Salmonella, Clostridium perfringens, Listeria monocytogenes, Lactobacillus, and E. coli O157:H7. These three active AMP produced obvious zones of inhibition on all 15 of the campylobacter isolates tested, regardless of the strain (Table 4). They also inhibited the other pathogens tested with the only exceptions being the lack of zone formation by the Cecropin-Magainin and RL-37 AMP on Lactobacillus helviticus and Clostridium perfringens 39.
MIC of C 12 K-2β 12 , Cecropin-Magainin, and RL-37 against 15 different C. jejuni, C. coli, and C. lari isolates are presented in Table 5. MIC for C 12 K-2β 12 ranged from 1.6 to 3.1 µg/mL with no discernable differences between strains of campylobacter. MIC for RL-37 varied from 1.6 to 6.3 µg/mL also with no obvious differentiation between campylobacter strains. The MIC for the Cecropin-Magainin hybrid were significantly higher, ranging from 12.5 to 100 µg/mL. No significant differences between campylobacter stains were observed in response to the Cecropin-Magainin AMP.

Discussion
AMP display tremendous chemical diversity in nature, causing difficulty in their classification [53,54]. Amino acid sequence, net charge, secondary structural motif, and the abundance of specific amino acids may all differ between types of AMP [55]. However, AMP also maintain some common features, and classification of AMP may be based on source, activity, structural characteristics, and amino acid-rich species [54]. For example, most AMP are cationic in nature with net charges ranging from +2 to +9. This gives an electrostatic advantage to the AMP in binding to the negatively charged bacterial membranes [56]. Many AMP are membrane-acting peptides and have an amphipathic structure with hydrophobic and hydrophilic regions which play an integral part in interactions with target bacterial membranes, leading to cell lysis through a variety of proposed mechanisms including the toroidal pore or wormhole model, the barrel-stave model, or the carpet-like model [57,58]. Indeed, some AMP are non-membrane-acting and do not cause cell lysis, but rather pass through the intact bacterial lipid bilayer through endocytosis and, once in the cytoplasm, inhibit nucleic acid biosynthesis, protease activity, or DNA replication [51]. A number of peptide databases exist and are available for mining [5,59,60]. We initially chose 11 AMP for evaluation representing a variety of different sources, modes of action, and activities frequently discussed in the literature ( Table 1). The selected AMP represent examples from insects, bacteria, amphibians, fish, mammals, and a case of de novo synthetic AMP. They also represent membrane-acting and non-membrane-acting peptides.
While very few reports exist in the literature specifically describing inhibition of campylobacters by AMP, C. jejuni has been reported to be highly susceptible to chicken host-defense peptides such as cathelicidin [61]. We therefore chose to evaluate RL-37, a 37-residue AMP of the cathelicidin family which is expressed in bone marrow of the rhesus monkey and utilizes an amphipathic α-helical structure to form pores and destabilize the membrane lipid bilayer of target bacteria. RL-37 is reported to inhibit Gram-negative organisms such as E. coli and Pseudomonas [48] and we found it to be quite efficacious against the Gram (−) campylobacters we tested, achieving MIC ranging from 1.6 to 6.3 µg/mL (Table 5). It was also reported to inhibit Gram (+) organisms such as Listeria monocytogenes and this was observed in our study as well (Table 4). We chose several additional α-helical AMP for evaluation including dermaseptin, NRC-Pleurocidin, Paracin I, and Temporin L. Dermaseptin, a 34-residue peptide isolated from frog skin (Phyllomedusa), demonstrates broad-spectrum lytic activity against Gram-positive and Gram-negative bacteria, yeast, and protozoa [43,62]. We found it to inhibit some, but not all, of our test isolates (Table 3). Since it did not inhibit both of our initial campylobacter isolates used for screening, we did not include it in further MIC evaluations. NRC-Pleurocidin, a 23-residue peptide isolated from the American plaice winter flounder has been reported to rapidly kill G (−) Pseudomonas aeruginosa and G (+) Methicillin-resistant S. aureus (MRSA) [45]. We found it capable of lysing all our test isolates except for C. perfringens 39. However, it was excluded from further analysis due to its moderate cytotoxicity. Paracin I is a 19-residue peptide isolated from catfish skin and has been reported to inhibit G (+) and G (−) organisms including E. coli and Salmonella [46] However, in the concentrations tested in our assay, Paracin I failed to inhibit any of our test isolates. Temporin L is a 13-residue peptide from European red frog skin and is among the shortest natural AMP known [49,63]. We found it to be efficacious against all our test organisms but because it is reported to be hemolytic and requires dimethylsulfoxide (DMSO) as a solvent rather than water, we did not think it practical for our purposes and did not include it in further analysis.
Some AMP contain a high content of certain amino acid residues such as proline, tryptophan, glycine, or histidine. These AMP are frequently extended with no regular structure [56]. Proline-rich AMP (Pr-AMPs) are cationic peptides which usually enter the cell through translocation of the inner membrane using a transporter-mediated uptake mechanism and do not lyse the cell. Pr-AMPs usually kill bacteria by targeting intracellular components. Pyrrhocoricin is a 20-amino acid residue Pr-AMP isolated from the European sap-sucking bug (Pyrrocoris apterus). Pyrrhocoricin interferes with protein biosynthesis in target bacteria by binding to the 70S ribosomal subunit and inhibiting heat-shock proteins such as DnaK, preventing protein folding [64]. In the concentrations tested in our assay, Pyrrhocoricin failed to inhibit any of our test isolates. Apidaecin 1B is 18-residue Pr-AMP isolated from honeybees. Apidaecin 1B is known to inhibit a large number of plantassociated bacteria and human pathogens via a bactericidal pathway without cell lysis [39]. Apidaecin 1B inhibited Salmonella enteritidis serovars Typhimurium and Heidelburg and E. coli O157:H7 in our assay, but did not kill Campylobacter jejuni, Clostridium perfringens, or Listeria monocytogenes isolates.
Bacteriocins are AMP produced by bacteria. Carnobacteriocin B2 is a Class II cationic bacteriocin which can form α-helices and lyse specific target bacteria [41]. Under the conditions of our spot-on-lawn assay, none of our target bacteria were affected by Carnobacteriocin B2. Dermcidin DCD is a 48-residue peptide isolated from human sweat glands. This anionic AMP forms oligomeric ion channels in lipid bilayers of specific target bacteria. This AMP was also unable to form any zones of inhibition against any of the bacteria utilized in our assay.
Despite the large number of natural AMP identified, isolated, and characterized, there are some inherent disadvantages associated with natural AMP. For example, the limited quantities of AMP isolated from natural sources, the uneconomic synthesis of longer AMP sequences, folding issues associated with natural AMP, short half-lives due to protease degradability, and difficulties in site-directed delivery [56]. This has led to a research emphasis on synthetic AMP in recent decades to overcome shortcomings of natural AMP. Several approaches have been employed for the design of improved synthetic AMP including template-based de novo synthesis based on natural AMP to improve activity and proteolytic stability (often through truncation), biophysical modeling in hydrophobic membrane environments, and virtual screening of peptide libraries to identify potential AMP sequences [56].
Various chemical modifications of AMP have been utilized to improve potency while decreasing cytotoxicity to mammalian cells, or improving peptide resistance to protease digestion, high temperatures, and low pH. Cecropin A (from the Cecropia moth) and Magainin 2 (from the African clawed frog, Xenopus Laevis) both individually exhibit potent antimicrobial effects against a broad spectrum of bacteria [65]. Shin and co-workers designed an engineered hybrid AMP composed of residues 1-8 of Cecropin A fused to residues 1-12 of Magainin 2 [66]. Cecropin-Magainin had greater antimicrobial activity against bacteria than either of the AMP individually while maintaining low toxicity. It has been suggested that the flexibility induced by Gly-Ile-Gly or Pro residues in the central part of Cecropin-Magainin may be important in the electrostatic and hydrophobic interactions of the N and C terminus with cell membrane surface and cell membrane, respectively [42]. We included Cecropin-Magainin in our spot-on-lawn assays and found it to inhibit all the bacterial species in our test, including the 15 Campylobacter sps. isolates, but not L. helviticus or C. perfringens 39 isolates.
Many naturally occurring AMP have rather long AA sequences which can limit their application as commercial drugs due to the high cost of producing them on an industrial scale [67]. Shorter synthetic AMP could potentially lower cost of synthesis and may be more resistant to proteolysis. Shorter synthetic AMP also could be less cumbersome to express in natural production vectors such as E. coli or Bacillus subtilis [68,69]. Oligomers of acylated lysines (OAKs) constitute a novel class of synthetic AMP made up of alternating amino acyl chains and cationic amino acids arranged in such a way as to create an optimal molecular charge and hydrophobicity for increased potency [70][71][72]. The synthetic OAK hexamer C 12 K-2β 12 exhibits inhibitory activity in vitro against Gram-negative bacteria such as the gastric pathogen Helicobacter pylori [40]. Because H. pylori has some structural characteristics in common with campylobacters, we chose to evaluate C 12 K-2β 12 in our spot-on-lawn assay. C 12 K-2β 12 was very active and inhibited growth of all target bacterial isolates in our assay. The bactericidal effects of C 12 K-2β 12 are exhibited by a dual mode of action. The OAK causes pore formation and cell lysis at higher concentrations while, at lower concentrations, it binds nucleic acids and proteins [55]. It demonstrated the lowest MIC of any of the AMP we tested against 15 isolates of Campylobacter representing jejuni, coli, and lari species, with a range of 1.6 to 3.1 ug/mL. C 12 K-2β 12 is only 8 amino acid residues in length and is reported to be stable at low pH and after exposure to extreme temperatures [40]. These are all important considerations for an AMP to be potentially administered to poultry. This AMP also inhibited other pathogenic bacterial species associated with poultry (Salmonella, C. perfringens). These characteristics make C 12 K-2β 12 a strong candidate for further investigation and potential development into a practical intervention for reduction of human foodborne pathogen populations associated with broiler chickens.
It is interesting that the AMP investigated in this study and found to be efficacious against Campylobacter spp. (NRC-13 Pleurocidin, RL-37, Temporin L, Cecropin-Magainin, Dermaseptin, and C 12 K-2β 12 ) are all cationic AMP hypothesized to form amphipathic α-helical structures to initiate pore formation in target bacteria and subsequent cell lysis. It is also interesting that Pyrrhocoricin, which is also a cationic, amphipathic α-helical AMP, did not inhibit Campylobacter spp. There is evidence in the literature suggesting that the selectivity and potency of a specific AMP is determined primarily by the chemical composition of the target membrane [1]. It has likewise been hypothesized that differences in membrane composition between different strains of bacteria are responsible for the diversity in the potency and selectivity exhibited by a particular AMP against different strains of bacteria [73].
One of the most important factors in reducing the burden of foodborne disease has been identified as development of effective control measures [74]. By providing novel alternatives to antibiotic usage in poultry, the overall impact of our long-term research goals will be a reduction in bacterial pathogens associated with chickens and safer products for human consumption. Our working hypothesis is that AMP that inhibit the growth of Campylobacter spp. can be identified and subsequently utilized to reduce the Campylobacter load among commercially produced chickens. Identifying non-hemolytic AMP, such as C 12 K-2β 12 , RL-37, and Cecropin-Magainin, capable of killing campylobacters in vitro is a positive first step toward this goal. An obvious second step will be producing efficacious AMP in an economical fashion. Microbiological production of AMP through fermentation is becoming a feasible and economic method of production. Recent studies have described the development of a Bacillus subtilis culture expressing chicken NK-2 peptide [69]. Oral delivery of the B. subtilis culture can protect against Eimeria acervuline infection in broiler chickens. We are currently working to similarly express effective AMP in a microbial vector for in vitro analysis and subsequent oral delivery to chickens to protect against colonization by Campylobacter spp.

Conclusions
Selected AMP produced obvious zones of inhibition against growth of C. jejuni isolates. These AMP included: NRC-13 Pleurocidin, RL-37, Temporin L, Cecropin A-Magainin 2 hybrid, Dermaseptin, and C 12 K-2β 12 . MIC for C 12 K-2β 12 ranged from 1.6 to 3.1 µg/mL with no discernable differences between strains of campylobacter. MIC for RL-37 varied from 1.6 to 6.3 µg/mL, also with no obvious differentiation between campylobacter strains. The MIC for Cecropin-Magainin were significantly higher, ranging from 12.5 to 100 µg/mL. No significant differences between campylobacter stains were observed in response to the Cecropin-Magainin AMP. The most efficacious of the AMP tested, the synthetic OAK C 12 K-2β 12 , is also heat-and acid-stable, making it an attractive subject for future study and potential in vivo delivery to poultry.