Antibacterial properties of the pituitary adenylate cyclase-activating polypeptide: A new human antimicrobial peptide

The Pituitary Adenylate Cyclase-Activating Polypeptide (PACAP), a polycationic, amphiphilic and helical neuropeptide, is well known for its neuroprotective actions and cell penetrating properties. In the present study, we evaluated the potent antibacterial property of PACAP38 and related analogs against various bacterial strains. Interestingly, PACAP38 and related analogs can inhibit the growth of various bacteria including Escherichia coli (JM109), Bacillus subtilis (PY79), and the pathogenic Burkholderia cenocepacia (J2315). Investigation of the mechanism of action suggested that a PACAP metabolite, identified as PACAP(9–38), might indeed be responsible for the observed PACAP38 antibacterial action. Surprisingly, PACAP(9–38), which does not induce haemolysis, exhibits an increased specificity toward Burkholderia cenocepacia J2315 compared to other tested bacteria. Finally, the predisposition of PACAP(9–38) to adopt a π-helix conformation rather than an α-helical conformation like PACAP38 could explain this gain in specificity. Overall, this study has revealed a new function for PACAP38 and related derivatives that can be added to its pleiotropic biological activities. This innovative study could therefore pave the way toward the development of new therapeutic agents against multiresistant bacteria, and more specifically the Burkholderia cenocepacia complex.


Introduction
Increased bacterial resistance to available antibiotics is a situation of global concern [1]. Based on their broad spectrum of activities and usually non-immunogenic action, antimicrobial peptides (AMPs) isolated from various species including plants, mammals, insects, and marine invertebrates [2], represent promising alternatives to traditional antibiotics. Natural antimicrobial peptides are relatively short polypeptides (fewer than 60 amino acid residues) that generally possess a positive net charge and amphipathic properties upon interaction with membranes [3]. Based on their structural (linear or helical) and biochemical properties (presence or absence of cysteine or over-representation of proline, arginine, tryptophan and histidine), antimicrobial peptides can be divided into at least five different groups [4]. Several companies have intensified their efforts to commercially develop AMPs, many of which are now in clinical trials [5]. However, their clinical use for human treatment remains limited to topical applications since AMPs are usually toxic or metabolically unstable when injected into the bloodstream [6]. Recently, the discovery of bacteria resistant to colistin, an antibiotic of last resort for many multidrug resistant bacteria was reported [7], further prompting the discovery of new treatments to fight a growing number of infections evading the latest generations of antibiotics. The pituitary adenylate cyclase-activating polypeptide is a 38 amino acid peptide (PACAP38) that belongs to the vasoactive intestinal polypeptide (VIP)-glucagon-growth hormone releasing factor-secretin superfamily [8]. This peptide, well known for its pleiotropic activities in human, mediates its biological activities through the activation of three G protein-coupled receptors, named PAC1, VPAC1, and VPAC2 [8]. Notably, PACAP38 and its analogs, mostly synthetic agonists, are currently regarded as promising candidates for the treatment of neurodegenerative disorders and brain trauma [9]. The recent discovery of intracellular PACAP38 receptors in the brain and the testis as well as the physicochemical characteristics of PACAP38, i.e. an extended amphipathic α-helix containing basic residues, prompted us to evaluate the propensity of PACAP38 to cross the plasma membrane in a receptor-independent manner [10]. We demonstrated the cell penetrating properties of PACAP38, which led to the development of a potent PACAP38-based cell penetrating peptide, [Arg 17 ]PACAP , with no cytotoxic effect on various eukaryotic cells including CHO-K1 [11]. Based on this unique property, and since most AMPs exhibit similar physicochemical and structural characteristics, we hypothesize that PACAP38 and its related derivatives can exert antimicrobial activities and therefore represent a new template for the development of a new class of innovative AMPs.
In this study, the antibacterial potential of PACAP38 as well as several selected derivatives was tested against various Gram-positive and Gram-negative bacteria. Our results demonstrate the potent antibacterial activity of PACAP38 against various bacterial strains, helping us identify a lead template, i.e. PACAP , that appears to be specific against Burkholderia cenocepacia, an epidemic pathogen of cystic fibrosis patients.

Materials
Fmoc-protected amino acids, Rink-amide AM resin and BOP (benzotriazol-1-yloxy)tris (dimethylamino)phosphonium hexafluorophosphate) reagents were purchased from Chem-Impex (Wood Dale, IL, USA). Solvents for solid phase peptide synthesis and purification were obtained from Tekniscience (Terrebonne, QC, CAN) whereas trifluoroacetic acid (TFA) was from PSIG (Montreal, QC, CAN). Na 125 I was purchased from Perkin Elmer (Montreal, QC, CAN). Other chemicals as well as cell culture media were obtained from Sigma-Aldrich (Mississauga, ON, CAN) and Fisher Scientific (Nepean, ON, CAN). The culture media for bacteria were purchased from BD Difco (Mississauga, ON, CAN).
Fmoc removal was achieved with 20% piperidine in DMF for 20 min. Peptide cleavage was achieved at room temperature using a mixture of TFA/ethanedithiol/phenol/water (92/2.5/3/ 2.5) for 3h. The diethyl ether-precipitated crude peptides were purified on a preparative RP-HPLC using a Phenomenex C 18 Gemini column and the collected fractions were analyzed concomitantly by analytical RP-HPLC, performed on a Phenomenex C 18 Jupiter column, and MALDI-TOF mass spectrometry using α-cyano-4-hydroxycinnamic acid as matrix (Voyager DE, Applied Biosystems). Fractions corresponding to the desired product and a purity higher than 95% were finally pooled and lyophilized.

Minimal inhibitory concentration (MIC) measurements
Assays were performed according to the CLSI microdilution method [12]. Briefly, the bacterial strains were inoculated and allowed to reach an OD 600 of 0.08-0.13 prior to dilution to 10 6 CFU/ mL in Mueller-Hinton broth (MHB). A bacterial suspension (50 μL) was then added to each well of a 96-well microplate containing serially diluted peptide solutions (50 μL in MHB) and incubated at 37˚C for 20h. A well containing sterilized medium was used as a sterility control and another well containing no added peptide served as a growth control. The MIC endpoint, expressed in μg/mL, represents the lowest concentration of PACAP38 or related analogs at which no bacterial growth was observed after the incubation time, as determined by OD measurements.

Membrane and cytoplasmic proteins extraction
Burkholderia cenocepacia strains J2315 and K56-2 membrane and cytoplasmic protein extraction was performed as previously described with some modifications [13]. Bacteria were grown in 8 mL of Tryptic Soy Broth (TSB) until they reached an OD 600 of 2 and then PACAP38 (200 μg/mL) was added to the cultures for 1h. The treated culture was then centrifuged at 7,000 x g for 15 min at 4˚C and the supernatant discarded. The resulting pellet was resuspended in 500 μL of cold Tris-HCl (10 mM; pH 8) containing 20% sucrose (w/v), frozen at -20˚C and then thawed to facilitate bacterial cells disruption. Following the addition of DNase (50 μg/mL), bacteria were lysed by FastPrep-24 (MPBio) using 0.1 mm glass beads (4.5 m/s, 60s). The suspension was then centrifuged at 10,000 x g for 30 min at 4˚C and the supernatant, containing the cytoplasmic proteins, recovered and kept at 4˚C until use. The beads were then mechanically removed and the remaining pellet was ultracentifuged in a sucrose gradient (50% and 70%) at 100,000 x g overnight at 4˚C in order to separate the inner (eluting between 20% and 40% of the sucrose gradient) and outer (eluting between 50% and 70% of the sucrose gradient) bacterial membranes. The resulting membrane fractions were collected in a minimum volume of water and kept at 4˚C until use. A similar procedure was used to extract membrane and cytoplasmic proteins from Bacillus subtilis strains PY79 and 2597 [14]. Briefly, Bacillus subtilis strains were grown in 8 mL of Tryptic Soy Broth (TSB) to an OD 600 of 2 and then PACAP38 (200 μg/ml) was added to the cultures for 1h before centrifugating at 8,000 x g for 10 min at 4˚C; finally, the pellet was washed twice with water. Bacterial cell lysis was then performed by FastPrep and the different protein extracts were isolated as described above, and kept at 4˚C until use.
Finally, the presence of PACAP or its metabolites in those different protein extracts was investigated by Matrix-Assisted Laser Desorption/Ionization Time-of-Flight (MALDI-TOF) mass spectrometry using α-cyano-4-hydroxycinnamic acid as the matrix.

Haemolytic activity
All experimental procedures were performed in accordance with regulations and ethical guidelines and approved by the institutional committee of the Institut National de la Recherche Scientifique-Institut Armand-Frappier. Blood from healthy human volunteers was collected into sodium citrate-buffered vacutainers. Blood donations were obtained from informed and consenting individuals according to institutionally approved procedures. Red blood cells were pelleted (700 x g; 10 min), washed three times with 9 volumes of PBS by centrifugation (700 x g; 10 min), and then resuspended in PBS. Cell suspensions (1 mL of final suspension) were incubated with different concentrations of peptide for 30 min or 3h, ranging from 10 −9 to 10 −5 M (4.5 ng/mL to 45 μg/mL for PACAP38 and 3.7 ng/mL to 37 μg/mL for PACAP (9-38)), at 37˚C with occasional mixing by inversion. PBS-incubated red blood cells were considered as negative control and maximum lysis of erythrocytes was obtained by incubating the cells with 0.1% (v/v) Triton X-100. At the end of the incubation period, cells were pelleted (700 x g, 10 min) and 100 μL of each supernatant was used to evaluate the release of hemoglobin by measuring the absorbance at 545 nm. Results were obtained from 3 independent experiments performed in triplicate.

Cell culture
Chinese hamster ovary (CHO) cells stably and individually expressing human receptor PAC1, VPAC1 or VPAC2 were grown in Ham's F12 medium supplemented with 10% of fetal bovine serum (FBS), 2 mM L-glutamine, 100 UI/mL of penicillin and streptomycin (P/S) and 400 μg/ mL of G418. The cell line was maintained at 37˚C in a humidified atmosphere of 5% CO 2 and passages were performed by trypsinization when cells reached about 80% confluence.

Radioligand binding assay
Acetylated PACAP27 was radioiodinated using the chloramine-T oxidation technique and purified on a Sep-Pak C 18 cartridge (Waters, Milford, MA, USA) [15]. Binding assays were performed using CHO cells stably transfected with the human PAC1, VPAC1 or VPAC2 receptor isoform. Cells were seeded at a density of 150,000 cells/well in 24-well plates. The next day, cells were washed with binding buffer (0.1% BSA, 25 mM Tris-HCl, 25 mM MgCl 2 , pH 7.4) for 10 min at room temperature. The solution was replaced with a fresh solution containing 0.05 nM of 125 I-Ac-PACAP27 and increasing concentrations of peptide (10 −14 to 10 -5 M). After 2h of incubation at room temperature, cells were washed twice with the binding buffer and then lysed with sodium hydroxide (0.1 M). Cell-bound radioactivity was quantified using a γcounter. Results were expressed as percentage of specific binding of 125 I-Ac-PACAP27. Nonspecific binding was determined in the presence of PACAP38 (10 -5 M).

Circular dichroism analysis
Circular dichroism (CD) spectra were recorded at room temperature from 200 to 250 nm, using a 1 mm optical path length with a 0.1 nm step, a 1 nm bandwidth, and an integration time of 4s on a Jasco J-815 Circular Dichroism (CD) Spectropolarimeter (Easton, MD, USA). Each spectrum represents the mean of three scans corrected for solvent contribution. A digital low-pass filter was used as a smoothing routine. Peptides were dissolved in 20 mM KH 2 PO 4 (pH 7.0) at a final concentration 200 μg/mL.

Molecular dynamic (MD) simulations of membrane-bound PACAP38 and PACAP(9-38)
MD simulations of the DPPC membrane-bound conformations of PACAP38 and PACAP(9-38) peptides were performed to characterize the secondary structure properties of the membrane-bound peptides. All simulations were performed under constant pressure, temperature and number of molecule (NPT) conditions, using the GROMACS simulation package v4.5.4. [16] Peptides were simulated using the GROMOS96 53A6 forcefield [17], while the lipids were described using the modified Berger force field parameters [18] for use with the GROMOS96 53A6 force field. The peptides were aligned parallel to the surface of the lipid bilayer and solvated with SPC water. Chloride ions were added to neutralize the charge of the system. The simulation system consisted of 2 x 64 DPPC lipids, one peptide (PACAP38 or PACAP(9-38)), and~7000 water molecules. The initial system box size was set to 6.4 nm x 6.4 nm x 9.5 nm. The systems were first energy minimized using the steepest descent method, followed by short equilibration runs before the long MD simulations. Two sets of simulations were performed for each of the two membrane-bound peptide systems (PACAP38 and PACAP(9-38)), corresponding to a total of 4 x 300 ns runs.
The temperature was set to 323 K using the Nose-Hoover thermostat with a coupling constant of 0.5 ps. The pressure was applied semi-isotropically and maintained using the Parinello-Rahman pressure coupling. All bonds and angles were constrained using the LINCS algorithm [19] and the particle mesh Ewald (PME) [20] method with a grid spacing of 0.16 nm was used for long-range interactions. The α-helical coordinates of micelle-bound PACAP38 (PDB: 2D2P), determined by NMR, were used as the starting conformation of the peptide. The structure of PACAP(9-38) was obtained by truncating PACAP38 to remove the first eight residues. The DPPC coordinates were taken from http://moose.bio.ucalgary.ca. Analysis of secondary structure propensity of the peptides along the simulation trajectory were performed using DSSP [21].
In vivo, PACAP exists in two isoforms of 27 and 38 amino acids that share structural, physicochemical and biological properties [8]. Both isoforms, containing between 5 and 11 basic residues, adopt an amphipathic helical conformation, starting around their ninth residue [9]. It is well known that the C-terminal domain of PACAP38 is involved in the stabilization of the αhelix [8]. In the present work, removal of the C-terminal domain of compound 7 produces a PACAP analog, i.e. [Tic 6 ]PACAP27 (8), which loses the ability to reduce bacterial growth (MIC > 160 μg/mL). Since both derivatives, i.e. [Tic 6 ]PACAP38 and [Tic 6 ]PACAP27, possess an extended C-terminal helix, the number of positively charged residues decorating those analogs (5 residues for compound 8 and 11 for 7) might be an essential trait to exert their antibacterial activity.
We next evaluated the specificity of PACAP activity against additional strains from the same species. While PACAP and its derivatives are highly active against strain J2315 of B. cenocepacia, they were inactive against four other strains belonging to the same species (Table 2). Additional strains of P. aeruginosa and E. coli were also evaluated with results mostly in accordance with the previously observed effects (Tables 1 and 2).

Insight into the PACAP38 mode of action
The known cell penetrating properties of 1 as well as its ability to interact with biological membranes prompted us to evaluate its propensity to interact with the bacterial membrane and to penetrate inside bacteria. To do so, 1 was incubated with various bacterial strains for 1h, allowing the peptide to enter the cells. No bacterial death was observed after this incubation period (data not shown). The bacterial cells were then fractionated into compartments and the presence of 1 or its metabolites in the outer/inner membranes or the cytoplasmic fraction was investigated by mass spectrometry (Table 3). As described above, PACAP38 can reduce the growth of B. subtilis strain PY79 but not of strain 2597 (Tables 1 and 2). Mass spectrometry analyses revealed that 1 (PACAP38) is found associated with the bacterial membrane in both strains but only present within the cytoplasmic fractions of PY79. Interestingly, a presumptive metabolite, identified as PACAP(9-38), was only observed associated with the sensitive PY79 strain. Furthermore, a similar pattern was also observed with the B. cenocepacia strains. While PACAP38 was observed on the outer membrane and within the cytoplasm of both B. cenocepacia strains, PACAP(9-38) was again only present within the cytoplasmic fractions of the sensitive J2315 strain. Hence, presence of the PACAP(9-38) metabolite in the cytoplasm seems to correlate with sensitivity to the inhibitory activity of PACAP38. The inability of 1 to reduce the growth homologous strains of Bacillus subtilis PY79 and Burkholderia cenocepacia J2315 might therefore point toward the existence of different proteases in these bacterial strains. Hence, while B. cenocepacia K56-2 and J2315 belong to the same clonal complex, they have several differences in their genomes and phenotypes [25]. Indeed, although there is only a draft genome of K-56-2 available (www. burkholderia.com), several differences in annotated proteases and peptidases are noted. Further experiments will be needed to confirm our assumption.

PACAP(9-38): A potentially selective antimicrobial peptide
To confirm the bactericidal activity of this PACAP38-derived metabolite, the analog was prepared by solid phase peptide synthesis. The pharmacological profile of this synthetic PACAP (9-38) was evaluated using three CHO cell lines respectively co-expressing the human PAC1, VPAC1 or VPAC2 receptors. As depicted in earlier publications, the N-terminal domain of PACAP plays an essential role for binding affinity and biological activity [8,23,26,27]. Hence, compared to the native PACAP38, PACAP(9-38) is characterized by a dramatic reduction of its binding affinity towards all PACAP receptors (Table 4), much like PACAP(11-38) (11). PACAP(9-38) was then tested against several PACAP38-sensitive bacterial strains. Unexpectedly, PACAP(9-38) was only able to reduce, with high efficiency (MIC: 19 μg/mL), the growth of strain J2315 of B. cenocepacia (Table 5), leaving unaffected several PACAP38-sensitive strains including P. putida (KT2440), E. coli (JM109), and B. subtilis (PY79). The percentage of haemolysis caused by various concentrations (4.5 ng/mL-45 μg/mL) of PACAP38 and (3.7 ng/mL-37 μg/mL) of PACAP(9-38) over 30 min or 3h at 37˚C is shown in Fig 1 At 37 or 45 μg/mL, PACAP38 and PACAP(9-38) do not exert a very significant haemolytic activity. While less prominent than P. aeruginosa-associated infection among cystic fibrosis patient, B. cenocepacia-related infection in such patients provokes a severe decline in lung function that might end up into a life-threatening systemic infection known as cepacia syndrome [28]. Treatment of such infection is extremely difficult as B. cenocepacia is highly resistant to multiple antibiotics or antibiotic combination [29]. With its apparent selectivity against B. cenocepacia J2315 and a potentially good therapeutic index, PACAP(9-38) represents a suitable candidate for the development of antimicrobial agents against cystic fibrosisassociated infection involving B. cenocepacia.

Relationships between the secondary structure and biological activity of PACAP38 and PACAP(9-38)
The secondary structure of PACAP38 and PACAP(9-38) was investigated by circular dichroism to ensure the conservation of the C-terminal α-helix [9,23]. As shown in Fig 2, PACAP38 and its truncated analog exhibit two negative minima at 208 and 222 nm in HFIP and 10% SDS (Fig 2B and 2C), characteristic of the presence of an α-helical structure. Conversely, in water, no structure was noticeable (Fig 2A).
Molecular dynamics (MD) simulations provide atomistic insights into structural properties of proteins. Here, we performed MD simulations of PACAP38 and PACAP(9-38) bound to the surface of a 1,2-dipalmitoylphosphatidylcholine (DPPC) lipid bilayer to characterize the secondary structural propensities for the two membrane-bound peptides. DPPC was chosen to mimic the zwitterionic phosphatidylethanolamine-rich membranes of gram-negative bacteria [30]. Two sets of 300 ns simulations were performed for each of the two peptides, corresponding to a cumulative simulation time of 1.2 μs. As previously demonstrated, this time scale provides a good approximation of peptide reorganization and structural behavior in solution [31]. A comparison of the secondary structural propensities of the two peptides confirms the predominant helical propensity in the two peptides (Fig 3). Interestingly, while PACAP38 showed a higher propensity for the formation of α-helical conformations relative to π-and 3 10 helices (Fig 3A and 3C), PACAP(9-38) consistently displayed a significantly higher propensity for πhelix conformations in the two independent simulations of the peptide (Fig 3B and 3D). This enhanced propensity of PACAP(9-38) to adopt π-helix conformations may contribute towards its higher selectivity against Burkholderia cenocepacia J2315. We note that while MD  (9-38). Statistical analysis was performed by using ANOVA, followed by a Dunnett's multiple comparison test, and differences were considered significant when ��� P < 0.001. https://doi.org/10.1371/journal.pone.0207366.g001 simulations provide important insights into the differences in the conformational properties of the two peptides, further experiments such as COSY and NOESY 2D NMR experiments would be necessary to confirm these observations.

Conclusions
We have demonstrated that the human neuropeptide PACAP38 possesses an antibacterial activity against a broad spectrum of bacteria. Notably, PACAP38 was able to inhibit the growth of the epidemic pathogen B. cenocepacia J2315 with a MIC of 10 μg/mL. Investigation about its mechanism of action suggested that a PACAP metabolite, identified as PACAP(9-38), might indeed be responsible for the observed PACAP38 antibacterial action on B. cenocepacia J2315 and B. subtilis PY79. Surprisingly, PACAP(9-38) exhibited an increased specificity toward B. cenocepacia J2315 compared to other tested bacteria, a phenomenon that might be related to  (9-38). The average percent α-helix, 3 10 and π helix as a function of residue number along the sequence are represented as red, green and blue lines, respectively, for two independent MD runs of PACAP38 (A, C) and PACAP9-38 (B, D).
https://doi.org/10.1371/journal.pone.0207366.g003 its ability to cross the bacterial membrane. Indeed, differences in the bacterial wall composition [32] might be responsible for the increased selectivity of PACAP(9-38), which was generally generated in situ within the bacterial cytoplasm following PACAP38 entry. Accordingly, it was observed that the uptake efficacy of PACAP within eukaryotic cell cytoplasm was dependent on the expression of cell surface glycosaminoglycans [33]. Another possibility relies on the predisposition of PACAP(9-38) to adopt π-helix conformations rather than α-helical conformations like PACAP38. While the mechanisms of action remain to be completely understood, we believe that PACAP38 and related derivatives can pave the way toward the development of new therapeutic agents against multidrug resistant bacteria, and more specifically the Burkholderia cepacia complex.