Proline-rich antimicrobial peptides show a long-lasting post-antibiotic effect on Enterobacteriaceae and Pseudomonas aeruginosa

Proline-rich antimicrobial peptides show a long-lasting post-antibiotic effect on... Abstract Background Proline-rich antimicrobial peptides (PrAMPs) represent a promising class of potential therapeutics to treat multiresistant infections. They inhibit bacterial protein translation at the 70S ribosome by either blocking the peptide-exit tunnel (oncocin type) or trapping release factors (apidaecin type). Objectives Besides direct concentration-dependent antibacterial effects, the post-antibiotic effect (PAE) is the second most important criterion of antimicrobial pharmacodynamics to be determined in vitro. Here, PAEs of 10 PrAMPs and three antibiotics against three Escherichia coli strains, Klebsiella pneumoniae ATCC 10031 and Pseudomonas aeruginosa ATCC 27853 were studied after 1 h of exposure. Methods A robust high-throughput screening to determine PAEs was established, i.e. liquid handling by a 96-channel pipetting system and continuous incubation and absorbance measurement in a microplate reader. Results Prolonged PAEs (≥4 h) were detected for all peptides at their MIC values against all strains; PAEs were even >10 h for Api88, Api137, Bac7(1-60) and A3-APO. The PAEs increased further at 4 × MIC. Aminoglycosides gentamicin and kanamycin usually showed lower PAEs (≤4 h) at MIC, but PAEs increased to > 10 h at 4 × MIC. Bacteriostatic chloramphenicol exhibited the shortest PAEs (<4 h). Conclusions The PAEs of PrAMPs studied against Enterobacteriaceae and P. aeruginosa for the first time were typically 4-fold stronger than for conventional antibiotics. Together with their fast and irreversible uptake by bacteria, the observed prolonged PAE of PrAMPs helps to explain their high in vivo efficacy despite unfavourable pharmacokinetics. Introduction Antibiotic-resistant Gram-negative bacteria represent a major health threat that is further worsened by the lack of novel antibiotic structures. Tigecycline (2005) and doripenem (2007), which are structurally related to tetracycline and carbapenem, respectively, were the last approved ‘novel’ antibiotics. Afterwards only a few ‘new’ substances representing modified β-lactams, aminoglycosides, quinolones or tetracyclines have been marketed. However, formulations overcoming resistance mechanisms by combining established antibiotics with new substances inhibiting bacterial enzymes that inactivate antibiotics, such as β-lactamase, have entered the market.1 Proline-rich antimicrobial peptides (PrAMPs) are pharmaceutically unused antimicrobial structures killing Gram-negative pathogens by at least two distinct intracellular novel mechanisms. PrAMPs are expressed in insects, such as apidaecin (Apis mellifera),2 drosocin (Drosophila melanogaster)3 and oncopeltus antimicrobial peptide 4 (Oncopeltus fasciatus),4 and in mammals (e.g. bactenecin Bac7 in bovine neutrophils).5 Additionally, PrAMPs like Chex1Arg20 and its dimer A3-APO have been designed.6,7 Mechanistically, the cationic PrAMPs interact with the negatively charged bacterial surface and translocate through the outer membrane.8 Especially long PrAMPs (>35 residues) disturb the bacterial membrane (at high concentrations) while translocating into the cytoplasm.9 Insect-derived PrAMPs are short (∼20 residues) and do not disrupt membranes, reducing harmful effects on mammalian cells. They pass the inner membrane via transporter-mediated uptake using protein SbmA10 and the efflux pump MdtM (drug/H+ antiporter).11 Interestingly, SbmA is present in phylogenetically distant species of Gram-negative bacteria, but absent in Pseudomonas aeruginosa.12 Still, optimized analogues of the short PrAMPs apidaecin and oncocin are active against P. aeruginosa without disrupting the bacterial membrane.13,14 This selective bacterial uptake usually prevents the translocation of short PrAMPs into mammalian cells, providing higher margins of safety that are further increased by their specific intracellular targets.15 Chaperone DnaK and the bacterial 70S ribosome are known targets of all PrAMPs, such as apidaecin analogue Api8816 and oncocin analogue Onc112.17 We classified PrAMPs by their different target mechanisms as apidaecin-type PrAMPs (e.g. Api88),18–20 which inhibit protein translation by trapping release factors at the ribosome,21 and oncocin-type PrAMPs (e.g. Onc112), which inhibit translation by blocking and destabilizing the initiation complex.22–25 Besides their low cytotoxicity, lack of haemolytic activity and good in vivo tolerance, the PrAMPs A3-APO, apidaecins, Bac7, drosocin and oncocin possess high efficacies in different infection models against Escherichia coli,16,26–28,Klebsiella pneumoniae,29,Salmonella enterica30 and Acinetobacter baumannii31 when administered intravenously,28 intraperitoneally,16,26,27,29,30 intramuscularly31,32 or subcutaneously.33 However, the pharmacokinetics does not explain the in vivo efficacy,34,35 at least in comparison with small-molecule antibiotics. Only continuous subcutaneous infusion of Api137 achieved stable plasma levels of ∼6 mg/L, which appeared reasonable to kill bacteria with corresponding MICs (D. Knappe, K. Adermann and R. Hoffmann, unpublished results). Remarkably, Craig36 noted already in 1991 that MICs and MBCs obtained by standard in vitro methods are insufficient in predicting in vivo antibiotic activity. Instead, further pharmacodynamic parameters have to be considered, such as post-antibiotic effect (PAE), post-antibiotic leucocyte enhancement (PALE) and minimal antibacterial concentration (MAC).37–39 PAE indicates the persistent growth of bacteria briefly exposed to antibiotics independently of host defence mechanisms. The PAE depends on different effects in vivo, such as persistent high drug concentrations at the site of infection, delayed recovery of enzyme and protein activity, changes in cell morphology, metabolism, cell receptors and ribosome density, prolonged growth and generation times, higher susceptibility to phagocytes and the influence of virulence factors like haemolysin.36–38 Pharmacodynamically, antibiotics are divided into three groups.37,39,40 The first group act mainly time dependently and are slowly bactericidal (e.g. β-lactams, vancomycin) with no or only low PAE. Their efficacy depends typically on the time they are present at plasma levels above MIC, demanding continuous or intermittent dosing.39 Group 2 are concentration-dependently bactericidal (e.g. aminoglycosides, fluoroquinolones) with Cmax as the limiting factor. As bactericidal activity and PAE increase with increasing dose, typically a single maximal dose is administered. The third group contains predominantly bacteriostatic antibiotics (e.g. macrolides) with moderate to prolonged PAE. Their action is concentration dependent rather than time dependent, with the AUC as the most important parameter.37,39,40 It is difficult to assign PrAMPs to any of these groups, although apidaecins appear to be described best as concentration-dependent (group 2) and oncocins as time-dependent antibiotics (group 1).35 In order to further classify short PrAMPs and to provide a better correlation of in vitro and in vivo data, the PAEs of 10 PrAMPs were determined in vitro in comparison with chloramphenicol, gentamicin and kanamycin (also ribosomal inhibitors) with known PAEs.41–44 Materials and methods Materials were obtained from the following manufacturers: standard nutrient agar I, nutrient broth (NB) and tryptic soy broth (TSB) from Carl Roth GmbH & Co. (Karlsruhe, Germany); Mueller–Hinton broth (MHB) from Sigma–Aldrich (Steinheim, Germany); and Gibco™ PBS (pH 7.4) from Thermo Fisher Scientific Inc. (Darmstadt, Germany). Peptides were synthesized in-house as described previously.3,4,16,34,35 Water was purified on a Purelab Ultra water purification system (electrical resistivity >182 kΩ·m; organic content <2 ppb). E. coli BW25113 was obtained from the Keio collection (GenoBase, http://ecoli.aist-nara.ac.jp).45E. coli MC4100 was kindly provided by Professor Dr Bernd Bukau (Heidelberg, Germany). E. coli ATCC 25922 (DSM1103), K. pneumoniae ATCC 10031 (DSM 681) and P. aeruginosa ATCC 27853 (DSM 1117) were purchased from Leibniz Institute DSMZ – German Collection of Microorganisms and Cell cultures (Braunschweig, Germany). Antimicrobial activity Peptides or antibiotics were plated in a 2-fold dilution series (50 μL/well) in MHB (5.75 g/L, 25% MHB) into 96-well F-bottom polystyrene plates (Greiner Bio-One GmbH, Frickenhausen, Germany, #655180). E. coli strains BW25113, MC4100 and ATCC 25922 and P. aeruginosa ATCC 27853 were grown overnight in nutrient broth and K. pneumoniae ATCC 10031 in TSB (4 mL, 200 rpm, 37 °C). Bacteria were diluted with 25% MHB to an OD600nm of ∼0.05 and incubated to reach a logarithmic growth rate between 1.5 and 4 h (Figure S1, available as Supplementary data at JAC Online). Bacterial suspensions were adjusted to 1.5 × 107 cells/mL and added to each well (50 μL) to obtain a final cell density of 7.5 × 106 cells/mL. After incubation (22 ± 2 h, 37 °C) the turbidity was measured at 595 nm with a Sunrise microplate reader (Tecan Deutschland GmbH, Crailsheim, Germany) for E. coli strains BW25113 and MC4100 or a SpectraMax 340 PC (Molecular Devices, Sunnyvale, CA, USA) for all other strains. MIC values were defined as the lowest peptide concentration inhibiting bacterial growth and were measured at least twice as triplicates on two different days. PAE Peptides or antibiotics were dissolved in water to final concentrations corresponding to 0.1 ×, 0.4 ×, 1 × and 4 × MIC into a sterile 96-well polystyrene V-bottom plate (3 μL/well; Greiner Bio-One GmbH, #651180). Water (3 μL/well) was used as negative control. Bacteria were grown and adjusted to 7.5 × 106 cells/mL, as described above, and added simultaneously to each well (147 μL) using the Steinbrenner Liquidator96® (Mettler Toledo, Gießen, Germany). All following steps were performed simultaneously using the Liquidator96®. After incubation (1 h, 37 °C, 600 rpm, Eppendorf Thermomixer), plates were centrifuged (1238 g, 10 min, 4 °C), the supernatant (130 μL/well) was removed and cells were suspended in fresh medium (room temperature, 150 μL/well). Plates were centrifuged again, the supernatant was removed and cells were again suspended in fresh medium (130 μL, 37 °C). An aliquot (100 μL) was transferred into anti-edge-effect polystyrene F-bottom plates (Eppendorf AG, Hamburg, Germany, #0030730.119) using 8 mL of water at the edge of the plate. The absorbance of the covered plate was recorded at 600 nm every 20 min after shaking (5 s) for 22 h (37 °C) using a Paradigm® microplate reader or a SpectraMax 340 PC (Molecular Devices). Cell viability was determined simultaneously on agar (37 °C; overnight) using the remaining cell suspension and appropriate dilutions (10 μL for each one-third of a 10 cm plate). The PAE was calculated as the difference between the time T needed for a cell culture incubated with an antibiotic substance to reach half-maximum absorbance of the control culture and the time C that the control culture needed to reach half-maximum absorbance of the stationary phase. The growth curves were fitted using a sigmoidal dose–response fit with variable slope (GraphPad Prism) according to equation (1). The software provides values for Bottom (minimum absorbance of the fit), Top (maximum absorbance), LogEC50 and HillSlope with the absorbance Y at the corresponding time X. The half-maximum absorbance (Ymax1/2) of the control was calculated using equation (2) and inserted in equation (3) to calculate the time for the treated culture (T) and the control (C). Equation (3) was transposed from equation (1).   Y= Bottom+Top-Bottom(1+10(log EC50-X*HillSlope)) (1)  Ymax⁡1/2= Abs(25% MHB)+Top(control)-Abs(25% MHB)2 (2)  X= - log10Top-BottomYmax1/2-Bottom-1-log EC50*HillSlopeHillSlope (3) For P. aeruginosa no clear stationary phase was observed and thus an automatic fit was impossible (Figure S2). Thus, Y was determined manually and Microsoft Excel was used to identify the closest absorbance of the sample to Ymax1/2 and the corresponding times C and T. Results Antimicrobial activity A prerequisite for determining PAEs of antibiotic substances is the MIC value determined under identical conditions. As in previous studies on PrAMPs,16–20,26 we used diluted (25%) MHB medium, which provides more reliable MIC values than full-strength MHB and allows much better prediction of the in vivo efficacy of PrAMPs. The MICs of most PrAMPs ranged from 2 to 8 mg/L for E. coli strains BW25113, MC4100 and ATCC 25922 and K. pneumoniae ATCC 10031 (Table 1). Only Bac7(1-60), drosocin Hyp5 and oncocin Onc72 were less active against some E. coli strains. The MIC values of gentamicin, kanamycin and chloramphenicol ranged from 0.25 to 2 mg/L. Considering the mean molecular weights for PrAMPs (2000 g/mol) and antibiotics (500 g/mol), similar molar antimicrobial activities were obtained against Enterobacteriaceae. Bac7(1-60) and A3-APO were equally active against P. aeruginosa ATCC 27853, whereas short PrAMPs were at least 16-fold less active (MIC ≥32 mg/L; Table 1). The better activity of long PrAMPs against P. aeruginosa probably relates to their effect on membranes, which is independent of transporter SbmA, which is absent in P. aeruginosa.9,12 The 4-fold higher MIC values of kanamycin and chloramphenicol are in accordance with the intrinsic resistance of P. aeruginosa to many antibiotics, whereas gentamicin remained highly active (MIC 1 mg/L). The MIC values of all three antibiotics determined here were in agreement with the literature.46 Table 1. Peptide sequences and MICs determined in 25% MHB Name  Peptide sequencea  MIC (mg/L)   E. coli BW25113  E. coli MC4100  E. coli ATCC 25922  K. pneumoniae ATCC 10031  P. aeruginosa ATCC 27853  Onc72  VDKPPYLPRPRPPROIYNO-NH2  32  32  32  2  128  Onc112  VDKPPYLPRPRPPRrIYNr-NH2  2  4  8  2  32  Api88  Gu-ONNRPVYIPRPRPPHPRL-NH2  2  2  4  4  32  Api137  Gu-ONNRPVYIPRPRPPHPRL-OH  2  2  2  1  64  Api795  Gu-OIOIORPVYOPRPRPPHPRL-OH  4  4  4  8  16  Bac7(1-60)  RRIRPRPPRLPRPRPRPLPFPRPGPRPIPRPL PFPRPGPRPIPRPLPFPRPGPRPIPRPL-OH  8  8  32  16  16  A3-APO  (Chex-RPDKPRPYLPRPRPPRPVR)2-Dab-NH2  8  8  16  16  16  Chex1Arg20  Chex-RPDKPRPYLPRPRPPRPVR-NH2  4  4  8  4  64  Drosocin Hyp5  GKPRxYSPRPTSHPRPIRV-OH  64  64  8  4  128  Pyrrhocoricin  VDKGSYLPRPTPPRPIYNRN-NH2  4  8  8  4  >128  Gentamicin  NA  0.25  0.5  0.5  0.5  1  Kanamycin  NA; mainly type A  0.25  0.5  0.5  0.5  64  Chloramphenicol  NA  2  2  2  1  128  Name  Peptide sequencea  MIC (mg/L)   E. coli BW25113  E. coli MC4100  E. coli ATCC 25922  K. pneumoniae ATCC 10031  P. aeruginosa ATCC 27853  Onc72  VDKPPYLPRPRPPROIYNO-NH2  32  32  32  2  128  Onc112  VDKPPYLPRPRPPRrIYNr-NH2  2  4  8  2  32  Api88  Gu-ONNRPVYIPRPRPPHPRL-NH2  2  2  4  4  32  Api137  Gu-ONNRPVYIPRPRPPHPRL-OH  2  2  2  1  64  Api795  Gu-OIOIORPVYOPRPRPPHPRL-OH  4  4  4  8  16  Bac7(1-60)  RRIRPRPPRLPRPRPRPLPFPRPGPRPIPRPL PFPRPGPRPIPRPLPFPRPGPRPIPRPL-OH  8  8  32  16  16  A3-APO  (Chex-RPDKPRPYLPRPRPPRPVR)2-Dab-NH2  8  8  16  16  16  Chex1Arg20  Chex-RPDKPRPYLPRPRPPRPVR-NH2  4  4  8  4  64  Drosocin Hyp5  GKPRxYSPRPTSHPRPIRV-OH  64  64  8  4  128  Pyrrhocoricin  VDKGSYLPRPTPPRPIYNRN-NH2  4  8  8  4  >128  Gentamicin  NA  0.25  0.5  0.5  0.5  1  Kanamycin  NA; mainly type A  0.25  0.5  0.5  0.5  64  Chloramphenicol  NA  2  2  2  1  128  NA, not applicable. a O, r, Gu, Chex, Dab and x denote l-ornithine, d-arginine, N,N,N´,N´-tetramethylguanidino, 1-cyclohexanecarboxylic acid, 2,4-diaminobutyric acid and trans-4-hydroxyproline, respectively. PAE Bacterial cultures were incubated with PrAMPs and antibiotic comparators at concentrations corresponding to 0.1 ×, 0.4 ×, 1 × and 4 × the respective MIC. Utilization of the 96-channel pipetting system enabled reproducible high-throughput analyses of 32 samples in triplicate on a 96-well microtitre plate (Figure 1). Besides similar curve shapes of the triplicates within one experiment, the PAE values were typically very reproducible among experiments on different days, e.g. 3.7 ± 0.6 h (at 0.4 × MIC) and 11.8 ± 2.4 h (at MIC) for Api137 and E. coli BW25113 (Figure 1b and Table S1). If bacterial growth remained <50% of the maximum growth of the control during the 14 h observation period, e.g. for Api137 at 4 × MIC (Figure 1b), the PAE was set to 14.0 h (strong PAEs). Figure 1. View largeDownload slide Growth curves. (a) E. coli (red) BW25113 (continuous line), MC4100 (dots) and ATCC 25922 (dashes), K. pneumoniae (blue) and P. aeruginosa (black) (7.5 × 105 cells/well; 100 µL) grown under standard conditions (1 h, 37 °C, 600 rpm; control). Shown is the average of at least five experiments performed in triplicate. (b) E. coli BW25113 (7.5 × 105 cells/well; 100 µL) grown under standard conditions in the absence (black) or presence of Api137 (green) at 0.1  × (dots), 0.4  × (short dashes), 1  × (long dashes) and 4  × (continuous line) MIC. One representative experiment in triplicate with each line representing the values of one well is shown. The black dashed line indicates the half-maximum absorbance of the untreated bacterial culture (control). Vertical lines represent the time period that the control (C) and the treated cultures needed to reach this density (T). Thus, Api137 concentrations of 0.1 ×, 0.4 ×, 1 × and 4 × MIC yielded PAEs (= T − C) of 0.4 ± 0.1, 3.3 ± 0.2, 10.6 ± 0.6 and >14.0 h, respectively. This figure appears in colour in the online version of JAC and in black and white in the print version of JAC. Figure 1. View largeDownload slide Growth curves. (a) E. coli (red) BW25113 (continuous line), MC4100 (dots) and ATCC 25922 (dashes), K. pneumoniae (blue) and P. aeruginosa (black) (7.5 × 105 cells/well; 100 µL) grown under standard conditions (1 h, 37 °C, 600 rpm; control). Shown is the average of at least five experiments performed in triplicate. (b) E. coli BW25113 (7.5 × 105 cells/well; 100 µL) grown under standard conditions in the absence (black) or presence of Api137 (green) at 0.1  × (dots), 0.4  × (short dashes), 1  × (long dashes) and 4  × (continuous line) MIC. One representative experiment in triplicate with each line representing the values of one well is shown. The black dashed line indicates the half-maximum absorbance of the untreated bacterial culture (control). Vertical lines represent the time period that the control (C) and the treated cultures needed to reach this density (T). Thus, Api137 concentrations of 0.1 ×, 0.4 ×, 1 × and 4 × MIC yielded PAEs (= T − C) of 0.4 ± 0.1, 3.3 ± 0.2, 10.6 ± 0.6 and >14.0 h, respectively. This figure appears in colour in the online version of JAC and in black and white in the print version of JAC. E. coli incubated with PrAMPs at 0.1 × MIC provided PAEs <4 h, e.g. 3.4 ± 0.9 and 3.7 ± 0.1 h for Onc72 against E. coli BW25113 and MC4100, respectively (Figure 2 and Table S1). The PAE typically remained <4 h when the peptide concentration was increased to 0.4 × MIC. Onc72 and A3-APO showed stronger effects against E. coli strains BW25113 (5.6 ± 0.8 and 4.6 ± 0.4 h, respectively) and MC4100 (9.2 ± 0.7 and 5.3 ± 0.6 h, respectively) and pyrrhocoricin against E. coli ATCC 25922 (4.4 ± 0.2 h). At MIC, the PAE increased for Onc72, Api88, Api137, Bac7(1-60) and Chex1Arg20 to >8 h for at least one of the three E. coli strains and for the longer peptides Bac7(1-60) and A3-APO to >12 h against E. coli strains BW25113 and MC4100. In contrast, gentamicin, kanamycin and chloramphenicol reached maximal PAEs of only 1.3 h, drosocin Hyp5 reached 4.7 h and pyrrhocoricin reached 6.2 h against all three E. coli strains. The highest PrAMP concentrations (4 × MIC) always resulted in PAEs >9 h, often even >12 h. The largest variations among the three E. coli strains were obtained for drosocin Hyp5 and pyrrhocoricin. The PAE of drosocin Hyp5 against E. coli ATCC 25922 was relatively strong at 4 × MIC (13.3 ± 4.7 h), but weak for BW25113 (4.1 ± 0.2 h) and MC4100 (2.9 ± 0.9 h). Pyrrhocoricin showed strong PAEs against E. coli ATCC 25922 (14.4 ± 3.0 h) and MC4100 (11.0 ± 2.4 h), but a weak PAE against BW25113 (3.0 ± 1.1 h). The PAEs determined for 4 × MIC were high for gentamicin (8–14 h), medium for kanamycin (6–10 h) and very low for chloramphenicol (2 h). All PrAMPs besides drosocin Hyp5 and pyrrhocoricin and all three small molecules showed similar PAEs against the E. coli strains. Figure 2. View large Download slide PAEs observed when treating E. coli strains BW25113 (top), MC4100 (middle) and ATCC 25922 (bottom) (7.5 × 106 cells/mL; 1 h, 37 °C, 600 rpm) with PrAMPs and antibiotic comparators at concentrations of 0.1 ×, 0.4 ×, 1 × and 4 × MIC (white, light grey, dark grey and black bars, respectively). Experiments were done twice in triplicate on different days. Hash signs indicate that none of the six individual curves reached half-maximum absorbance of the control within 22 h, whereas dollar signs indicate that this threshold was not reached in one to five replicates. For details see Table S1. Figure 2. View large Download slide PAEs observed when treating E. coli strains BW25113 (top), MC4100 (middle) and ATCC 25922 (bottom) (7.5 × 106 cells/mL; 1 h, 37 °C, 600 rpm) with PrAMPs and antibiotic comparators at concentrations of 0.1 ×, 0.4 ×, 1 × and 4 × MIC (white, light grey, dark grey and black bars, respectively). Experiments were done twice in triplicate on different days. Hash signs indicate that none of the six individual curves reached half-maximum absorbance of the control within 22 h, whereas dollar signs indicate that this threshold was not reached in one to five replicates. For details see Table S1. The PAEs of PrAMPs and standard antibiotics against K. pneumoniae ATCC 10031 were consistently <3 h at 0.1 × MIC, with the highest value obtained for Api88 (2.6 ± 0.3 h) (Figure 3 and Table S1). Fourfold higher concentrations increased the PAEs of Api88 (7.7 ± 0.4 h), Api137 (4.2 ± 0.6 h) and A3-APO (4.4 ± 0.9 h) significantly, whereas the PAEs of all other compounds remained <4 h. Interestingly, at MIC the PAE increased for Api88 (18.8 ± 1.2 h), Api795 (13.4 ± 4.7 h), Bac7(1-60) (18.2 ± 3.3 h) and A3-APO (13.1 ± 2.7 h) to >12 h, but for all other peptides remained below 12 h, with drosocin Hyp5 showing the weakest PAE of 4.2 ± 0.5 h. The PAEs of the three small-molecule antibiotics were even lower (1.3–3.6 h at MIC). At high concentrations (4 × MIC) most PrAMPs exhibited strong PAEs against K. pneumoniae (≥18 h); only drosocin Hyp5 (10.3 ± 0.6 h), pyrrhocoricin (9.8 ± 0.5 h) and Chex1Arg 20 (6.6 ± 0.7 h) showed lower values. Note that at these high concentrations both aminoglycosides (gentamicin and kanamycin) exhibited as strong PAEs as the best PrAMPs (≥19.5 h), whereas the PAE of chloramphenicol remained at 3.5 ± 0.2 h. Figure 3. View large Download slide PAEs obtained by incubating K. pneumoniae ATCC 10031 (7.5 × 106 cells/mL; 1 h, 37 °C, 600 rpm) with PrAMPs and antibiotic comparators at concentrations of 0.1 ×, 0.4 ×, 1 × and 4 × MIC (white, light grey, dark grey and black bars, respectively). Experiments were done twice in triplicate on different days. The PAE was calculated as T − C, with C being the time the control needed to reach half-maximum absorbance and T the time necessary for the treated sample to reach the half-maximum absorbance of the control. Hash signs indicate that none of the six individual curves reached half-maximum absorbance of the control within 22 h, whereas dollar signs indicate that this threshold was not reached in one to five replicates. For details see Table S1. Figure 3. View large Download slide PAEs obtained by incubating K. pneumoniae ATCC 10031 (7.5 × 106 cells/mL; 1 h, 37 °C, 600 rpm) with PrAMPs and antibiotic comparators at concentrations of 0.1 ×, 0.4 ×, 1 × and 4 × MIC (white, light grey, dark grey and black bars, respectively). Experiments were done twice in triplicate on different days. The PAE was calculated as T − C, with C being the time the control needed to reach half-maximum absorbance and T the time necessary for the treated sample to reach the half-maximum absorbance of the control. Hash signs indicate that none of the six individual curves reached half-maximum absorbance of the control within 22 h, whereas dollar signs indicate that this threshold was not reached in one to five replicates. For details see Table S1. When P. aeruginosa ATCC 27853 was inhibited with small-molecule antibiotics at concentrations up to MIC, the PAEs remained <4 h (Figure 4 and Table S1). This was also true for PrAMPs at sub-MIC, but at MIC the PAEs of PrAMPs ranged from 4.3 ± 0.3 h (Api137) to 8.7 ± 1.2 h (A3-APO). At 4 × MIC the PAEs were all >10 h, with Onc112 showing the best PAE (16.8 ± 1.3 h). Drosocin Hyp5 and pyrrhocoricin were not tested as they were virtually inactive against P. aeruginosa. Gentamicin and kanamycin exhibited lower PAEs than the PrAMPs (7.8 ± 0.7 and 8.9 ± 1.1 h). Chloramphenicol showed the weakest PAE of all compounds (2.1 ± 0.7 h). Figure 4. View large Download slide PAEs obtained for P. aeruginosa ATCC 27853 (7.5 × 106 cells/mL; 1 h, 37 °C, 600 rpm) when incubated with PrAMPs and antibiotic comparators at concentrations of 0.1 ×, 0.4 ×, 1 × and 4 × MIC (white, light grey, dark grey and black bars, respectively). Experiments were done at least twice in triplicate on different days. The PAE was calculated as T − C with C being the time the control needed to reach half-maximum absorbance and T the time necessary for the treated sample to reach the half-maximum absorbance of the control. Dollar signs indicate that one to five individual curves did not reach the half-maximum absorbance of the control within 22 h. For details see Table S1. Figure 4. View large Download slide PAEs obtained for P. aeruginosa ATCC 27853 (7.5 × 106 cells/mL; 1 h, 37 °C, 600 rpm) when incubated with PrAMPs and antibiotic comparators at concentrations of 0.1 ×, 0.4 ×, 1 × and 4 × MIC (white, light grey, dark grey and black bars, respectively). Experiments were done at least twice in triplicate on different days. The PAE was calculated as T − C with C being the time the control needed to reach half-maximum absorbance and T the time necessary for the treated sample to reach the half-maximum absorbance of the control. Dollar signs indicate that one to five individual curves did not reach the half-maximum absorbance of the control within 22 h. For details see Table S1. PAE and viable cell count The viable cell counts of E. coli and K. pneumoniae usually decreased at MIC by at least 1 log unit within an incubation period of 1 h and even further for conditions providing long PAEs (Figures S3, S4 and S5). The viable count of the P. aeruginosa control was much lower (2.7 ± 1.1 × 105 cfu/mL) than the mean of all other strains (5.1 ± 2.2 × 106 cfu/mL). Chloramphenicol always showed bacteriostatic activities. Drosocin Hyp5 and pyrrhocoricin were also bacteriostatic except for E. coli ATCC 25922 (P. aeruginosa not tested). Interestingly, Chex1Arg20 was bacteriostatic solely against K. pneumoniae, for which it also exhibited the lowest PAE. Gentamicin and kanamycin were always bactericidal at 4 × MIC. Strong bactericidal activities were observed for the long PrAMPs Bac7(1-60) and A3-APO, which can be explained by their dual mode of action, i.e. intracellular inhibition at MIC and increasing membrane permeability at peptide concentrations above the MIC.9 Interestingly, the bactericidal effects obtained for PrAMPs against E. coli ATCC 25922 and P. aeruginosa exceeded the effects of aminoglycosides. Discussion There are several studies about PAEs, but standardized protocols are missing. Thus, the bacteria used, bacterial and antibiotic concentrations, exposure times and methods for drug removal vary in the literature, with only a few publications relying on comparable conditions (for a detailed discussion see the Supplementary Discussion available as Supplementary data at JAC Online).37,E. coli ATCC 25922 is commonly used in PAE studies and the MIC values of gentamicin (0.5 mg/L) and chloramphenicol (2 mg/L) measured here for this strain were comparable to previous studies.42–44 However, the PAEs of 8.0 ± 0.7 h (4 × MIC) and 2.6 ± 0.6 h (1 × MIC) obtained here for gentamicin lasted much longer than in previous reports, i.e. 1.1 h (4 × MIC) and 0.5 h (1 × MIC),44 but were close to the 2 and 4 h reported for gentamicin at 5 × MIC.42,47 The PAE of 1 h reported for bacteriostatic chloramphenicol (5 × MIC) was close to the 2.0 ± 0.5 h measured here at 4 × MIC.42 These differences are most likely related to different laboratory routines, especially the different cell densities. Our sensitive method, which enables the determination of PAEs in ∼20 h, relied on the cell density recommended for MIC assays, i.e. 7.5 × 105 cells/mL. In contrast, the literature relies on much higher cell densities of 2 × 107 to 2 × 108 cells/mL, which results also in higher MIC values when measured for these conditions.13,48 Importantly, the high-throughput method described here enabled determination of the PAE for four different concentrations of 10 PrAMPs and three antibiotic comparators with high reproducibility. Unexpectedly, the PAEs of several PrAMPs determined at 0.4 × MIC exceeded 4 h, i.e. Onc72 (≥5.6 h) and A3-APO (≥4.6 h) against E. coli MC4100 and BW25113, pyrrhocoricin (4.4 h) against E. coli ATCC 25922 and Api88 (7.7 h) and Api137 (4.2 h) even against K. pneumoniae ATCC 10031. The PAEs determined for PrAMPs at MIC lasted typically for >4 h, with the highest PAEs (≥10 h) obtained against at least one E. coli strain and K. pneumoniae ATCC 10031 for Api88, Api137, Bac7(1-60) and A3-APO and for Chex1Arg20 against E. coli BW25113 and MC4100. The effects seen for small-molecule antibiotics at MIC were considerably weaker. Only gentamicin exceeded 4 h against E. coli MC4100. As expected, the PAEs further increased for higher peptide concentrations and almost all substances showed values >10 h against all strains at 4 × MIC, which is partially related to the bactericidal effect, i.e. reduced viable cell numbers after the incubation period. The same was true for gentamicin and kanamycin, possessing PAEs comparable to PrAMPs, whereas the PAEs of chloramphenicol remained <4 h, indicating its bacteriostatic properties. As a statistical analysis considering all tested substances and bacterial strains would provide a very large data set, we compared the apidaecin and oncocin derivatives, which possess similar structures and properties, as one group with the three antibiotics (small molecules) as a second group. The apidaecin/oncocin group showed significantly higher PAEs than the group of antibiotics at 0.4 × and 1 × MIC against all strains, underlining their high therapeutic potential (Figure S6). Drosocin Hyp5 and pyrrhocoricin exhibited the lowest PAEs among the PrAMPs, comparable to gentamicin and kanamycin. The PAEs of the other PrAMPs increased in the following order: oncocins < apidaecins < Bac7(1-60) = A3-APO. The long-lasting PAEs of the two long PrAMPs can be explained by their additional effects on bacterial membranes.9,12 Interestingly, some PrAMPs reduced the viable cell counts after an incubation period of 1 h at 4 × MIC more than the bactericidal aminoglycosides. Considering the long PAEs of PrAMPs observed here at MIC and sub-MIC and the bactericidal activity indicated by decreasing viable cell counts, PrAMPs probably belong to group 2 of the pharmacodynamic classification scheme, i.e. antibiotics with concentration-dependent bactericidal action and prolonged PAEs, such as aminoglycosides, with Cmax as limiting factor.39 Although the strong PAEs were unexpected, a recent study on bacterial uptake perfectly correlates to this observation. Apidaecins and oncocins bind rapidly to E. coli and translocate into the cytoplasm within minutes, as indicated by labelled peptides using flow cytometry and confirmed by unlabelled peptides using liquid chromatography.11,48,49 The authors could show that internalized Api137 remains in the cells without leaking into fresh medium using conditions close to the PAE assay applied here.48 In contrast, Onc112 was detected in the fresh medium. However, the authors argued that Onc112 might be released mostly from non-viable cells. These uptake studies and the current PAE data can nicely explain the apparent discrepancy between high in vivo efficacy and relatively low plasma levels.29,34,35 Presumably, PrAMPs enter bacteria rapidly in the host and inhibit the 70S ribosomes, as indicated by the strong PAE, even at MIC and partially also at sub-MIC concentrations. Thus, bacterial growth is stopped, allowing the host immune system to eradicate the pathogens before they start to grow again. It should be noted that Laszlo Otvos’s group has provided compelling evidence that their designer peptide A3-APO stimulates the mouse immune system as a third mode of action.50 In this respect, the efficacy of A3-APO might be best described as an artificial host defence peptide that relies mostly on immunomodulation, as demonstrated in wound and systemic infection models for Propionibacterium acnes, Staphylococcus aureus and Proteus mirabilis.7,51,52 Most recently, they showed that pre-infection treatment with transdermally administered A3-APO protects mice against intraperitoneal infections with A. baumannii.53 Apidaecins and oncocins have not been studied in vivo for immunomodulatory effects yet. In vitro, neither type of PrAMP showed immunomodulatory effects on unstimulated and LPS-stimulated murine dendritic cells or macrophages.54 In contrast, reduction of LPS-induced TNF-α release of human macrophages and monocytes by apidaecins indicates a mild anti-inflammatory effect.55 Independently, Tavano et al.56 reported for apidaecin 1b a differential immunomodulation of human macrophages, dendritic cells and macrophages. Additionally, A3-APO increases the levels of anti-inflammatory cytokines IL-4 and IL-10 in human peripheral blood mononuclear cells, possibly indicating a reduction in inflammation at the site of an infection.52 The PAEs presented here provide an additional hypothesis besides immunomodulatory effects that can explain the good in vivo efficacies of PrAMPs, which are not wholly expected, in different infection models despite the fast clearance rates reported for apidaecins, oncocins, A3-APO and Chex1Arg20 in mice. This highlights again that MIC values determined for AMPs in vitro cannot be simply used to predict in vivo efficacies, as often assumed in the literature. Instead MIC values should be seen as one important criterion among other parameters to be considered. Acknowledgements We thank Tina Goldbach and Dr Martina Bluhm for technical support in peptide synthesis and the Federal Ministry of Education and Research (BMBF), the European Fund for Regional Structure Development (EFRE) and the German Research Foundation (DFG) for financial support. Funding This work was supported by the Federal Ministry of Education and Research (BMBF; grant number 01GU1104A), the European Fund for Regional Structure Development (EFRE; European Union and Free State of Saxony; grant numbers 100105139 and 100127675) and the German Research Foundation (DFG; grant number INST 268/289-1). Transparency declarations D. K. was a part-time co-worker of AMP Therapeutics (Leipzig, Germany). R. H. is a co-founder of AMP Therapeutics and a member of its scientific advisory board. L. H.: none to declare. Author contributions L. H. conducted the experiments. L. H. and D. K. designed the experiments. L. H., D. K. and R. H. wrote the manuscript. Supplementary data Figures S1 to S6, Table S1 and Supplementary Discussion are available as Supplementary data at JAC Online. References 1 Falagas ME, Mavroudis AD, Vardakas KZ. The antibiotic pipeline for multi-drug resistant Gram-negative bacteria: what can we expect? Expert Rev Anti Infect Ther  2016; 14: 747– 63. Google Scholar CrossRef Search ADS   2 Casteels P, Ampe C, Jacobs F et al.   Apidaecins: antibacterial peptides from honeybees. EMBO J  1989; 8: 2387– 91. 3 Knappe D, Cassone M, Nollmann FI et al.   Hydroxyproline substitutions stabilize non-glycosylated drosocin against serum proteases without challenging its antibacterial activity. Protein Pept Lett  2014; 21: 321– 9. Google Scholar CrossRef Search ADS   4 Knappe D, Piantavigna S, Hansen A et al.   Oncocin (VDKPPYLPRPRPPRRIYNR-NH2): a novel antibacterial peptide optimized against Gram-negative human pathogens. J Med Chem  2010; 53: 5240– 7. Google Scholar CrossRef Search ADS   5 Giacometti A, Cirioni O, Del Prete MS et al.   In vitro effect on Cryptosporidium parvum of short-term exposure to cathelicidin peptides. J Antimicrob Chemother  2003; 51: 843– 7. Google Scholar CrossRef Search ADS   6 Otvos LJr, Wade JD, Lin F et al.   Designer antibacterial peptides kill fluoroquinolone-resistant clinical isolates. J Med Chem  2005; 48: 5349– 59. Google Scholar CrossRef Search ADS   7 Ostorhazi E, Voros E, Nemes-Nikodem E et al.   Rapid systemic and local treatments with the antibacterial peptide dimer A3-APO and its monomeric metabolite eliminate bacteria and reduce inflammation in intradermal lesions infected with Propionibacterium acnes and meticillin-resistant Staphylococcus aureus. Int J Antimicrob Agents  2013; 42: 537– 43. Google Scholar CrossRef Search ADS   8 Scocchi M, Mardirossian M, Runti G et al.   Non-membrane permeabilizing modes of action of antimicrobial peptides on bacteria. Curr Top Med Chem  2016; 16: 76– 88. Google Scholar CrossRef Search ADS   9 Podda E, Benincasa M, Pacor S et al.   Dual mode of action of Bac7, a proline-rich antibacterial peptide. Biochim Biophys Acta  2006; 1760: 1732– 40. Google Scholar CrossRef Search ADS   10 Mattiuzzo M, Bandiera A, Gennaro R et al.   Role of the Escherichia coli SbmA in the antimicrobial activity of proline-rich peptides. Mol Microbiol  2007; 66: 151– 63. Google Scholar CrossRef Search ADS   11 Krizsan A, Knappe D, Hoffmann R. Influence of the yjiL-mdtM gene cluster on the antibacterial activity of proline-rich antimicrobial peptides overcoming Escherichia coli resistance induced by the missing SbmA transporter system. Antimicrob Agents Chemother  2015; 59: 5992– 8. Google Scholar CrossRef Search ADS   12 Graf M, Mardirossian M, Nguyen F et al.   Proline-rich antimicrobial peptides targeting protein synthesis. Nat Prod Rep  2017; 34: 702– 11. Google Scholar CrossRef Search ADS   13 Bluhm ME, Schneider VA, Schafer I et al.   N-terminal Ile-Orn- and Trp-Orn-motif repeats enhance membrane interaction and increase the antimicrobial activity of apidaecins against Pseudomonas aeruginosa. Front Cell Dev Biol  2016; 4: 39. Google Scholar CrossRef Search ADS   14 Knappe D, Ruden S, Langanke S et al.   Optimization of oncocin for antibacterial activity using a SPOT synthesis approach: extending the pathogen spectrum to Staphylococcus aureus. Amino Acids  2016; 48: 269– 80. Google Scholar CrossRef Search ADS   15 Otvos LJr. The short proline-rich antibacterial peptide family. Cell Mol Life Sci  2002; 59: 1138– 50. Google Scholar CrossRef Search ADS   16 Czihal P, Knappe D, Fritsche S et al.   Api88 is a novel antibacterial designer peptide to treat systemic infections with multidrug-resistant Gram-negative pathogens. ACS Chem Biol  2012; 7: 1281– 91. Google Scholar CrossRef Search ADS   17 Knappe D, Kabankov N, Hoffmann R. Bactericidal oncocin derivatives with superior serum stabilities. Int J Antimicrob Agents  2011; 37: 166– 70. Google Scholar CrossRef Search ADS   18 Volke D, Krizsan A, Berthold N et al.   Identification of Api88 binding partners in Escherichia coli using a photoaffinity-cross-link strategy and label-free quantification. J Proteome Res  2015; 14: 3274– 83. Google Scholar CrossRef Search ADS   19 Krizsan A, Volke D, Weinert S et al.   Insect-derived proline-rich antimicrobial peptides kill bacteria by inhibiting bacterial protein translation at the 70 S ribosome. Angew Chem Int Ed Engl  2014; 53: 12236– 9. Google Scholar CrossRef Search ADS   20 Krizsan A, Prahl C, Goldbach T et al.   Short proline-rich antimicrobial peptides inhibit either the bacterial 70S ribosome or the assembly of its large 50S subunit. Chembiochem  2015; 16: 2304– 8. Google Scholar CrossRef Search ADS   21 Florin T, Maracci C, Graf M et al.   An antimicrobial peptide that inhibits translation by trapping release factors on the ribosome. Nat Struct Mol Biol  2017; 24: 752– 7. Google Scholar CrossRef Search ADS   22 Gagnon MG, Roy RN, Lomakin IB et al.   Structures of proline-rich peptides bound to the ribosome reveal a common mechanism of protein synthesis inhibition. Nucleic Acids Res  2016; 44: 2439– 50. Google Scholar CrossRef Search ADS   23 Roy RN, Lomakin IB, Gagnon MG et al.   The mechanism of inhibition of protein synthesis by the proline-rich peptide oncocin. Nat Struct Mol Biol  2015; 22: 466– 9. Google Scholar CrossRef Search ADS   24 Seefeldt AC, Nguyen F, Antunes S et al.   The proline-rich antimicrobial peptide Onc112 inhibits translation by blocking and destabilizing the initiation complex. Nat Struct Mol Biol  2015; 22: 470– 5. Google Scholar CrossRef Search ADS   25 Seefeldt AC, Graf M, Perebaskine N et al.   Structure of the mammalian antimicrobial peptide Bac7(1-16) bound within the exit tunnel of a bacterial ribosome. Nucleic Acids Res  2016; 44: 2429– 38. Google Scholar CrossRef Search ADS   26 Knappe D, Fritsche S, Alber G et al.   Oncocin derivative Onc72 is highly active against Escherichia coli in a systemic septicaemia infection mouse model. J Antimicrob Chemother  2012; 67: 2445– 51. Google Scholar CrossRef Search ADS   27 Szabo D, Ostorhazi E, Binas A et al.   The designer proline-rich antibacterial peptide A3-APO is effective against systemic Escherichia coli infections in different mouse models. Int J Antimicrob Agents  2010; 35: 357– 61. Google Scholar CrossRef Search ADS   28 Hoffmann R, Bulet P, Urge L et al.   Range of activity and metabolic stability of synthetic antibacterial glycopeptides from insects. Biochim Biophys Acta  1999; 1426: 459– 67. Google Scholar CrossRef Search ADS   29 Ostorhazi E, Nemes-Nikodem E, Knappe D et al.   In vivo activity of optimized apidaecin and oncocin peptides against a multiresistant, KPC-producing Klebsiella pneumoniae strain. Protein Pept Lett  2014; 21: 368– 73. Google Scholar CrossRef Search ADS   30 Benincasa M, Pelillo C, Zorzet S et al.   The proline-rich peptide Bac7(1-35) reduces mortality from Salmonella typhimurium in a mouse model of infection. BMC Microbiol  2010; 10: 178. Google Scholar CrossRef Search ADS   31 Ostorhazi E, Rozgonyi F, Sztodola A et al.   Preclinical advantages of intramuscularly administered peptide A3-APO over existing therapies in Acinetobacter baumannii wound infections. J Antimicrob Chemother  2010; 65: 2416– 22. Google Scholar CrossRef Search ADS   32 Ostorhazi E, Rozgonyi F, Szabo D et al.   Intramuscularly administered peptide A3-APO is effective against carbapenem-resistant Acinetobacter baumannii in mouse models of systemic infections. Biopolymers  2011; 96: 126– 9. Google Scholar CrossRef Search ADS   33 Knappe D, Adermann K, Hoffmann R. Oncocin Onc72 is efficacious against antibiotic-susceptible Klebsiella pneumoniae ATCC 43816 in a murine thigh infection model. Biopolymers  2015; 104: 707– 11. Google Scholar CrossRef Search ADS   34 Schmidt R, Ostorhazi E, Wende E et al.   Pharmacokinetics and in vivo efficacy of optimized oncocin derivatives. J Antimicrob Chemother  2016; 71: 1003– 11. Google Scholar CrossRef Search ADS   35 Schmidt R, Knappe D, Wende E et al.   In vivo efficacy and pharmacokinetics of optimized apidaecin analogs. Front Chem  2017; 5: 15. Google Scholar CrossRef Search ADS   36 Craig WA. The postantibiotic effect. Clin Microbiol Newslett  1991; 13: 121– 4. Google Scholar CrossRef Search ADS   37 MacKenzie FM, Gould IM. The post-antibiotic effect. J Antimicrob Chemother  1993; 32: 519– 37. Google Scholar CrossRef Search ADS   38 Lowdin E, Odenholt-Tornqvist I, Bengtsson S et al.   A new method to determine postantibiotic effect and effects of subinhibitory antibiotic concentrations. Antimicrob Agents Chemother  1993; 37: 2200– 5. Google Scholar CrossRef Search ADS   39 Levison ME, Levison JH. Pharmacokinetics and pharmacodynamics of antibacterial agents. Infect Dis Clin North Am  2009; 23: 791– 815, vii. Google Scholar CrossRef Search ADS   40 Craig W. Pharmacodynamics of antimicrobial agents as a basis for determining dosage regimens. Eur J Clin Microbiol Infect Dis  1993; 12 Suppl 1: S6– 8. Google Scholar CrossRef Search ADS   41 Tenson T, Mankin A. Antibiotics and the ribosome. Mol Microbiol  2006; 59: 1664– 77. Google Scholar CrossRef Search ADS   42 Stubbings WJ, Bostock JM, Ingham E et al.   Assessment of a microplate method for determining the post-antibiotic effect in Staphylococcus aureus and Escherichia coli. J Antimicrob Chemother  2004; 54: 139– 43. Google Scholar CrossRef Search ADS   43 Zhanel GG, Karlowsky JA, Davidson RJ et al.   Effect of pooled human cerebrospinal fluid on the postantibiotic effects of cefotaxime, ciprofloxacin, and gentamicin against Escherichia coli. Antimicrob Agents Chemother  1992; 36: 1136– 9. Google Scholar CrossRef Search ADS   44 Dominguez MC, de La Rosa M, Borobio MV. Application of a spectrophotometric method for the determination of post-antibiotic effect and comparison with viable counts in agar. J Antimicrob Chemother  2001; 47: 391– 8. Google Scholar CrossRef Search ADS   45 Baba T, Ara T, Hasegawa M et al.   Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the Keio collection. Mol Syst Biol  2006; 2: 2006.0008. Google Scholar CrossRef Search ADS   46 Fass RJ, Barnishan J. Minimal inhibitory concentrations of 34 antimicrobial agents for control strains Escherichia coli ATCC 25922 and Pseudomonas aeruginosa ATCC 27853. Antimicrob Agents Chemother  1979; 16: 622– 4. Google Scholar CrossRef Search ADS   47 Stubbings W, Bostock J, Ingham E et al.   Mechanisms of the post-antibiotic effects induced by rifampicin and gentamicin in Escherichia coli. J Antimicrob Chemother  2006; 58: 444– 8. Google Scholar CrossRef Search ADS   48 Holfeld L, Hoffmann R, Knappe D. Correlating uptake and activity of proline-rich antimicrobial peptides in Escherichia coli. Anal Bioanal Chem  2017; 409: 5581– 92. Google Scholar CrossRef Search ADS   49 Berthold N, Hoffmann R. Cellular uptake of apidaecin 1b and related analogs in Gram-negative bacteria reveals novel antibacterial mechanism for proline-rich antimicrobial peptides. Protein Pept Lett  2014; 21: 391– 8. Google Scholar CrossRef Search ADS   50 Otvos L. Racing on the wrong track. Front Chem  2017; 5: 42. Google Scholar CrossRef Search ADS   51 Otvos LJr, Ostorhazi E. Therapeutic utility of antibacterial peptides in wound healing. Expert Rev Anti Infect Ther  2015; 13: 871– 81. Google Scholar CrossRef Search ADS   52 Ostorhazi E, Holub MC, Rozgonyi F et al.   Broad-spectrum antimicrobial efficacy of peptide A3-APO in mouse models of multidrug-resistant wound and lung infections cannot be explained by in vitro activity against the pathogens involved. Int J Antimicrob Agents  2011; 37: 480– 4. Google Scholar CrossRef Search ADS   53 Ostorhazi E, Horvath A, Szabo D et al.   Transdermally administered proline-arginine-rich host defense peptides show systemic efficacy in a lethal mouse bacteremia model. Amino Acids  2017; 49: 1647– 51. Google Scholar CrossRef Search ADS   54 Fritsche S, Knappe D, Berthold N et al.   Absence of in vitro innate immunomodulation by insect-derived short proline-rich antimicrobial peptides points to direct antibacterial action in vivo. J Pept Sci  2012; 18: 599– 608. Google Scholar CrossRef Search ADS   55 Keitel U, Schilling E, Knappe D et al.   Effect of antimicrobial peptides from Apis mellifera hemolymph and its optimized version Api88 on biological activities of human monocytes and mast cells. Innate Immun  2013; 19: 355– 67. Google Scholar CrossRef Search ADS   56 Tavano R, Segat D, Gobbo M et al.   The honeybee antimicrobial peptide apidaecin differentially immunomodulates human macrophages, monocytes and dendritic cells. J Innate Immun  2011; 3: 614– 22. Google Scholar CrossRef Search ADS   © The Author 2017. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please email: journals.permissions@oup.com. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Antimicrobial Chemotherapy Oxford University Press

Proline-rich antimicrobial peptides show a long-lasting post-antibiotic effect on Enterobacteriaceae and Pseudomonas aeruginosa

Loading next page...
 
/lp/ou_press/proline-rich-antimicrobial-peptides-show-a-long-lasting-post-00RPi9fvDe
Publisher
Oxford University Press
ISSN
0305-7453
eISSN
1460-2091
D.O.I.
10.1093/jac/dkx482
Publisher site
See Article on Publisher Site

Abstract

Abstract Background Proline-rich antimicrobial peptides (PrAMPs) represent a promising class of potential therapeutics to treat multiresistant infections. They inhibit bacterial protein translation at the 70S ribosome by either blocking the peptide-exit tunnel (oncocin type) or trapping release factors (apidaecin type). Objectives Besides direct concentration-dependent antibacterial effects, the post-antibiotic effect (PAE) is the second most important criterion of antimicrobial pharmacodynamics to be determined in vitro. Here, PAEs of 10 PrAMPs and three antibiotics against three Escherichia coli strains, Klebsiella pneumoniae ATCC 10031 and Pseudomonas aeruginosa ATCC 27853 were studied after 1 h of exposure. Methods A robust high-throughput screening to determine PAEs was established, i.e. liquid handling by a 96-channel pipetting system and continuous incubation and absorbance measurement in a microplate reader. Results Prolonged PAEs (≥4 h) were detected for all peptides at their MIC values against all strains; PAEs were even >10 h for Api88, Api137, Bac7(1-60) and A3-APO. The PAEs increased further at 4 × MIC. Aminoglycosides gentamicin and kanamycin usually showed lower PAEs (≤4 h) at MIC, but PAEs increased to > 10 h at 4 × MIC. Bacteriostatic chloramphenicol exhibited the shortest PAEs (<4 h). Conclusions The PAEs of PrAMPs studied against Enterobacteriaceae and P. aeruginosa for the first time were typically 4-fold stronger than for conventional antibiotics. Together with their fast and irreversible uptake by bacteria, the observed prolonged PAE of PrAMPs helps to explain their high in vivo efficacy despite unfavourable pharmacokinetics. Introduction Antibiotic-resistant Gram-negative bacteria represent a major health threat that is further worsened by the lack of novel antibiotic structures. Tigecycline (2005) and doripenem (2007), which are structurally related to tetracycline and carbapenem, respectively, were the last approved ‘novel’ antibiotics. Afterwards only a few ‘new’ substances representing modified β-lactams, aminoglycosides, quinolones or tetracyclines have been marketed. However, formulations overcoming resistance mechanisms by combining established antibiotics with new substances inhibiting bacterial enzymes that inactivate antibiotics, such as β-lactamase, have entered the market.1 Proline-rich antimicrobial peptides (PrAMPs) are pharmaceutically unused antimicrobial structures killing Gram-negative pathogens by at least two distinct intracellular novel mechanisms. PrAMPs are expressed in insects, such as apidaecin (Apis mellifera),2 drosocin (Drosophila melanogaster)3 and oncopeltus antimicrobial peptide 4 (Oncopeltus fasciatus),4 and in mammals (e.g. bactenecin Bac7 in bovine neutrophils).5 Additionally, PrAMPs like Chex1Arg20 and its dimer A3-APO have been designed.6,7 Mechanistically, the cationic PrAMPs interact with the negatively charged bacterial surface and translocate through the outer membrane.8 Especially long PrAMPs (>35 residues) disturb the bacterial membrane (at high concentrations) while translocating into the cytoplasm.9 Insect-derived PrAMPs are short (∼20 residues) and do not disrupt membranes, reducing harmful effects on mammalian cells. They pass the inner membrane via transporter-mediated uptake using protein SbmA10 and the efflux pump MdtM (drug/H+ antiporter).11 Interestingly, SbmA is present in phylogenetically distant species of Gram-negative bacteria, but absent in Pseudomonas aeruginosa.12 Still, optimized analogues of the short PrAMPs apidaecin and oncocin are active against P. aeruginosa without disrupting the bacterial membrane.13,14 This selective bacterial uptake usually prevents the translocation of short PrAMPs into mammalian cells, providing higher margins of safety that are further increased by their specific intracellular targets.15 Chaperone DnaK and the bacterial 70S ribosome are known targets of all PrAMPs, such as apidaecin analogue Api8816 and oncocin analogue Onc112.17 We classified PrAMPs by their different target mechanisms as apidaecin-type PrAMPs (e.g. Api88),18–20 which inhibit protein translation by trapping release factors at the ribosome,21 and oncocin-type PrAMPs (e.g. Onc112), which inhibit translation by blocking and destabilizing the initiation complex.22–25 Besides their low cytotoxicity, lack of haemolytic activity and good in vivo tolerance, the PrAMPs A3-APO, apidaecins, Bac7, drosocin and oncocin possess high efficacies in different infection models against Escherichia coli,16,26–28,Klebsiella pneumoniae,29,Salmonella enterica30 and Acinetobacter baumannii31 when administered intravenously,28 intraperitoneally,16,26,27,29,30 intramuscularly31,32 or subcutaneously.33 However, the pharmacokinetics does not explain the in vivo efficacy,34,35 at least in comparison with small-molecule antibiotics. Only continuous subcutaneous infusion of Api137 achieved stable plasma levels of ∼6 mg/L, which appeared reasonable to kill bacteria with corresponding MICs (D. Knappe, K. Adermann and R. Hoffmann, unpublished results). Remarkably, Craig36 noted already in 1991 that MICs and MBCs obtained by standard in vitro methods are insufficient in predicting in vivo antibiotic activity. Instead, further pharmacodynamic parameters have to be considered, such as post-antibiotic effect (PAE), post-antibiotic leucocyte enhancement (PALE) and minimal antibacterial concentration (MAC).37–39 PAE indicates the persistent growth of bacteria briefly exposed to antibiotics independently of host defence mechanisms. The PAE depends on different effects in vivo, such as persistent high drug concentrations at the site of infection, delayed recovery of enzyme and protein activity, changes in cell morphology, metabolism, cell receptors and ribosome density, prolonged growth and generation times, higher susceptibility to phagocytes and the influence of virulence factors like haemolysin.36–38 Pharmacodynamically, antibiotics are divided into three groups.37,39,40 The first group act mainly time dependently and are slowly bactericidal (e.g. β-lactams, vancomycin) with no or only low PAE. Their efficacy depends typically on the time they are present at plasma levels above MIC, demanding continuous or intermittent dosing.39 Group 2 are concentration-dependently bactericidal (e.g. aminoglycosides, fluoroquinolones) with Cmax as the limiting factor. As bactericidal activity and PAE increase with increasing dose, typically a single maximal dose is administered. The third group contains predominantly bacteriostatic antibiotics (e.g. macrolides) with moderate to prolonged PAE. Their action is concentration dependent rather than time dependent, with the AUC as the most important parameter.37,39,40 It is difficult to assign PrAMPs to any of these groups, although apidaecins appear to be described best as concentration-dependent (group 2) and oncocins as time-dependent antibiotics (group 1).35 In order to further classify short PrAMPs and to provide a better correlation of in vitro and in vivo data, the PAEs of 10 PrAMPs were determined in vitro in comparison with chloramphenicol, gentamicin and kanamycin (also ribosomal inhibitors) with known PAEs.41–44 Materials and methods Materials were obtained from the following manufacturers: standard nutrient agar I, nutrient broth (NB) and tryptic soy broth (TSB) from Carl Roth GmbH & Co. (Karlsruhe, Germany); Mueller–Hinton broth (MHB) from Sigma–Aldrich (Steinheim, Germany); and Gibco™ PBS (pH 7.4) from Thermo Fisher Scientific Inc. (Darmstadt, Germany). Peptides were synthesized in-house as described previously.3,4,16,34,35 Water was purified on a Purelab Ultra water purification system (electrical resistivity >182 kΩ·m; organic content <2 ppb). E. coli BW25113 was obtained from the Keio collection (GenoBase, http://ecoli.aist-nara.ac.jp).45E. coli MC4100 was kindly provided by Professor Dr Bernd Bukau (Heidelberg, Germany). E. coli ATCC 25922 (DSM1103), K. pneumoniae ATCC 10031 (DSM 681) and P. aeruginosa ATCC 27853 (DSM 1117) were purchased from Leibniz Institute DSMZ – German Collection of Microorganisms and Cell cultures (Braunschweig, Germany). Antimicrobial activity Peptides or antibiotics were plated in a 2-fold dilution series (50 μL/well) in MHB (5.75 g/L, 25% MHB) into 96-well F-bottom polystyrene plates (Greiner Bio-One GmbH, Frickenhausen, Germany, #655180). E. coli strains BW25113, MC4100 and ATCC 25922 and P. aeruginosa ATCC 27853 were grown overnight in nutrient broth and K. pneumoniae ATCC 10031 in TSB (4 mL, 200 rpm, 37 °C). Bacteria were diluted with 25% MHB to an OD600nm of ∼0.05 and incubated to reach a logarithmic growth rate between 1.5 and 4 h (Figure S1, available as Supplementary data at JAC Online). Bacterial suspensions were adjusted to 1.5 × 107 cells/mL and added to each well (50 μL) to obtain a final cell density of 7.5 × 106 cells/mL. After incubation (22 ± 2 h, 37 °C) the turbidity was measured at 595 nm with a Sunrise microplate reader (Tecan Deutschland GmbH, Crailsheim, Germany) for E. coli strains BW25113 and MC4100 or a SpectraMax 340 PC (Molecular Devices, Sunnyvale, CA, USA) for all other strains. MIC values were defined as the lowest peptide concentration inhibiting bacterial growth and were measured at least twice as triplicates on two different days. PAE Peptides or antibiotics were dissolved in water to final concentrations corresponding to 0.1 ×, 0.4 ×, 1 × and 4 × MIC into a sterile 96-well polystyrene V-bottom plate (3 μL/well; Greiner Bio-One GmbH, #651180). Water (3 μL/well) was used as negative control. Bacteria were grown and adjusted to 7.5 × 106 cells/mL, as described above, and added simultaneously to each well (147 μL) using the Steinbrenner Liquidator96® (Mettler Toledo, Gießen, Germany). All following steps were performed simultaneously using the Liquidator96®. After incubation (1 h, 37 °C, 600 rpm, Eppendorf Thermomixer), plates were centrifuged (1238 g, 10 min, 4 °C), the supernatant (130 μL/well) was removed and cells were suspended in fresh medium (room temperature, 150 μL/well). Plates were centrifuged again, the supernatant was removed and cells were again suspended in fresh medium (130 μL, 37 °C). An aliquot (100 μL) was transferred into anti-edge-effect polystyrene F-bottom plates (Eppendorf AG, Hamburg, Germany, #0030730.119) using 8 mL of water at the edge of the plate. The absorbance of the covered plate was recorded at 600 nm every 20 min after shaking (5 s) for 22 h (37 °C) using a Paradigm® microplate reader or a SpectraMax 340 PC (Molecular Devices). Cell viability was determined simultaneously on agar (37 °C; overnight) using the remaining cell suspension and appropriate dilutions (10 μL for each one-third of a 10 cm plate). The PAE was calculated as the difference between the time T needed for a cell culture incubated with an antibiotic substance to reach half-maximum absorbance of the control culture and the time C that the control culture needed to reach half-maximum absorbance of the stationary phase. The growth curves were fitted using a sigmoidal dose–response fit with variable slope (GraphPad Prism) according to equation (1). The software provides values for Bottom (minimum absorbance of the fit), Top (maximum absorbance), LogEC50 and HillSlope with the absorbance Y at the corresponding time X. The half-maximum absorbance (Ymax1/2) of the control was calculated using equation (2) and inserted in equation (3) to calculate the time for the treated culture (T) and the control (C). Equation (3) was transposed from equation (1).   Y= Bottom+Top-Bottom(1+10(log EC50-X*HillSlope)) (1)  Ymax⁡1/2= Abs(25% MHB)+Top(control)-Abs(25% MHB)2 (2)  X= - log10Top-BottomYmax1/2-Bottom-1-log EC50*HillSlopeHillSlope (3) For P. aeruginosa no clear stationary phase was observed and thus an automatic fit was impossible (Figure S2). Thus, Y was determined manually and Microsoft Excel was used to identify the closest absorbance of the sample to Ymax1/2 and the corresponding times C and T. Results Antimicrobial activity A prerequisite for determining PAEs of antibiotic substances is the MIC value determined under identical conditions. As in previous studies on PrAMPs,16–20,26 we used diluted (25%) MHB medium, which provides more reliable MIC values than full-strength MHB and allows much better prediction of the in vivo efficacy of PrAMPs. The MICs of most PrAMPs ranged from 2 to 8 mg/L for E. coli strains BW25113, MC4100 and ATCC 25922 and K. pneumoniae ATCC 10031 (Table 1). Only Bac7(1-60), drosocin Hyp5 and oncocin Onc72 were less active against some E. coli strains. The MIC values of gentamicin, kanamycin and chloramphenicol ranged from 0.25 to 2 mg/L. Considering the mean molecular weights for PrAMPs (2000 g/mol) and antibiotics (500 g/mol), similar molar antimicrobial activities were obtained against Enterobacteriaceae. Bac7(1-60) and A3-APO were equally active against P. aeruginosa ATCC 27853, whereas short PrAMPs were at least 16-fold less active (MIC ≥32 mg/L; Table 1). The better activity of long PrAMPs against P. aeruginosa probably relates to their effect on membranes, which is independent of transporter SbmA, which is absent in P. aeruginosa.9,12 The 4-fold higher MIC values of kanamycin and chloramphenicol are in accordance with the intrinsic resistance of P. aeruginosa to many antibiotics, whereas gentamicin remained highly active (MIC 1 mg/L). The MIC values of all three antibiotics determined here were in agreement with the literature.46 Table 1. Peptide sequences and MICs determined in 25% MHB Name  Peptide sequencea  MIC (mg/L)   E. coli BW25113  E. coli MC4100  E. coli ATCC 25922  K. pneumoniae ATCC 10031  P. aeruginosa ATCC 27853  Onc72  VDKPPYLPRPRPPROIYNO-NH2  32  32  32  2  128  Onc112  VDKPPYLPRPRPPRrIYNr-NH2  2  4  8  2  32  Api88  Gu-ONNRPVYIPRPRPPHPRL-NH2  2  2  4  4  32  Api137  Gu-ONNRPVYIPRPRPPHPRL-OH  2  2  2  1  64  Api795  Gu-OIOIORPVYOPRPRPPHPRL-OH  4  4  4  8  16  Bac7(1-60)  RRIRPRPPRLPRPRPRPLPFPRPGPRPIPRPL PFPRPGPRPIPRPLPFPRPGPRPIPRPL-OH  8  8  32  16  16  A3-APO  (Chex-RPDKPRPYLPRPRPPRPVR)2-Dab-NH2  8  8  16  16  16  Chex1Arg20  Chex-RPDKPRPYLPRPRPPRPVR-NH2  4  4  8  4  64  Drosocin Hyp5  GKPRxYSPRPTSHPRPIRV-OH  64  64  8  4  128  Pyrrhocoricin  VDKGSYLPRPTPPRPIYNRN-NH2  4  8  8  4  >128  Gentamicin  NA  0.25  0.5  0.5  0.5  1  Kanamycin  NA; mainly type A  0.25  0.5  0.5  0.5  64  Chloramphenicol  NA  2  2  2  1  128  Name  Peptide sequencea  MIC (mg/L)   E. coli BW25113  E. coli MC4100  E. coli ATCC 25922  K. pneumoniae ATCC 10031  P. aeruginosa ATCC 27853  Onc72  VDKPPYLPRPRPPROIYNO-NH2  32  32  32  2  128  Onc112  VDKPPYLPRPRPPRrIYNr-NH2  2  4  8  2  32  Api88  Gu-ONNRPVYIPRPRPPHPRL-NH2  2  2  4  4  32  Api137  Gu-ONNRPVYIPRPRPPHPRL-OH  2  2  2  1  64  Api795  Gu-OIOIORPVYOPRPRPPHPRL-OH  4  4  4  8  16  Bac7(1-60)  RRIRPRPPRLPRPRPRPLPFPRPGPRPIPRPL PFPRPGPRPIPRPLPFPRPGPRPIPRPL-OH  8  8  32  16  16  A3-APO  (Chex-RPDKPRPYLPRPRPPRPVR)2-Dab-NH2  8  8  16  16  16  Chex1Arg20  Chex-RPDKPRPYLPRPRPPRPVR-NH2  4  4  8  4  64  Drosocin Hyp5  GKPRxYSPRPTSHPRPIRV-OH  64  64  8  4  128  Pyrrhocoricin  VDKGSYLPRPTPPRPIYNRN-NH2  4  8  8  4  >128  Gentamicin  NA  0.25  0.5  0.5  0.5  1  Kanamycin  NA; mainly type A  0.25  0.5  0.5  0.5  64  Chloramphenicol  NA  2  2  2  1  128  NA, not applicable. a O, r, Gu, Chex, Dab and x denote l-ornithine, d-arginine, N,N,N´,N´-tetramethylguanidino, 1-cyclohexanecarboxylic acid, 2,4-diaminobutyric acid and trans-4-hydroxyproline, respectively. PAE Bacterial cultures were incubated with PrAMPs and antibiotic comparators at concentrations corresponding to 0.1 ×, 0.4 ×, 1 × and 4 × the respective MIC. Utilization of the 96-channel pipetting system enabled reproducible high-throughput analyses of 32 samples in triplicate on a 96-well microtitre plate (Figure 1). Besides similar curve shapes of the triplicates within one experiment, the PAE values were typically very reproducible among experiments on different days, e.g. 3.7 ± 0.6 h (at 0.4 × MIC) and 11.8 ± 2.4 h (at MIC) for Api137 and E. coli BW25113 (Figure 1b and Table S1). If bacterial growth remained <50% of the maximum growth of the control during the 14 h observation period, e.g. for Api137 at 4 × MIC (Figure 1b), the PAE was set to 14.0 h (strong PAEs). Figure 1. View largeDownload slide Growth curves. (a) E. coli (red) BW25113 (continuous line), MC4100 (dots) and ATCC 25922 (dashes), K. pneumoniae (blue) and P. aeruginosa (black) (7.5 × 105 cells/well; 100 µL) grown under standard conditions (1 h, 37 °C, 600 rpm; control). Shown is the average of at least five experiments performed in triplicate. (b) E. coli BW25113 (7.5 × 105 cells/well; 100 µL) grown under standard conditions in the absence (black) or presence of Api137 (green) at 0.1  × (dots), 0.4  × (short dashes), 1  × (long dashes) and 4  × (continuous line) MIC. One representative experiment in triplicate with each line representing the values of one well is shown. The black dashed line indicates the half-maximum absorbance of the untreated bacterial culture (control). Vertical lines represent the time period that the control (C) and the treated cultures needed to reach this density (T). Thus, Api137 concentrations of 0.1 ×, 0.4 ×, 1 × and 4 × MIC yielded PAEs (= T − C) of 0.4 ± 0.1, 3.3 ± 0.2, 10.6 ± 0.6 and >14.0 h, respectively. This figure appears in colour in the online version of JAC and in black and white in the print version of JAC. Figure 1. View largeDownload slide Growth curves. (a) E. coli (red) BW25113 (continuous line), MC4100 (dots) and ATCC 25922 (dashes), K. pneumoniae (blue) and P. aeruginosa (black) (7.5 × 105 cells/well; 100 µL) grown under standard conditions (1 h, 37 °C, 600 rpm; control). Shown is the average of at least five experiments performed in triplicate. (b) E. coli BW25113 (7.5 × 105 cells/well; 100 µL) grown under standard conditions in the absence (black) or presence of Api137 (green) at 0.1  × (dots), 0.4  × (short dashes), 1  × (long dashes) and 4  × (continuous line) MIC. One representative experiment in triplicate with each line representing the values of one well is shown. The black dashed line indicates the half-maximum absorbance of the untreated bacterial culture (control). Vertical lines represent the time period that the control (C) and the treated cultures needed to reach this density (T). Thus, Api137 concentrations of 0.1 ×, 0.4 ×, 1 × and 4 × MIC yielded PAEs (= T − C) of 0.4 ± 0.1, 3.3 ± 0.2, 10.6 ± 0.6 and >14.0 h, respectively. This figure appears in colour in the online version of JAC and in black and white in the print version of JAC. E. coli incubated with PrAMPs at 0.1 × MIC provided PAEs <4 h, e.g. 3.4 ± 0.9 and 3.7 ± 0.1 h for Onc72 against E. coli BW25113 and MC4100, respectively (Figure 2 and Table S1). The PAE typically remained <4 h when the peptide concentration was increased to 0.4 × MIC. Onc72 and A3-APO showed stronger effects against E. coli strains BW25113 (5.6 ± 0.8 and 4.6 ± 0.4 h, respectively) and MC4100 (9.2 ± 0.7 and 5.3 ± 0.6 h, respectively) and pyrrhocoricin against E. coli ATCC 25922 (4.4 ± 0.2 h). At MIC, the PAE increased for Onc72, Api88, Api137, Bac7(1-60) and Chex1Arg20 to >8 h for at least one of the three E. coli strains and for the longer peptides Bac7(1-60) and A3-APO to >12 h against E. coli strains BW25113 and MC4100. In contrast, gentamicin, kanamycin and chloramphenicol reached maximal PAEs of only 1.3 h, drosocin Hyp5 reached 4.7 h and pyrrhocoricin reached 6.2 h against all three E. coli strains. The highest PrAMP concentrations (4 × MIC) always resulted in PAEs >9 h, often even >12 h. The largest variations among the three E. coli strains were obtained for drosocin Hyp5 and pyrrhocoricin. The PAE of drosocin Hyp5 against E. coli ATCC 25922 was relatively strong at 4 × MIC (13.3 ± 4.7 h), but weak for BW25113 (4.1 ± 0.2 h) and MC4100 (2.9 ± 0.9 h). Pyrrhocoricin showed strong PAEs against E. coli ATCC 25922 (14.4 ± 3.0 h) and MC4100 (11.0 ± 2.4 h), but a weak PAE against BW25113 (3.0 ± 1.1 h). The PAEs determined for 4 × MIC were high for gentamicin (8–14 h), medium for kanamycin (6–10 h) and very low for chloramphenicol (2 h). All PrAMPs besides drosocin Hyp5 and pyrrhocoricin and all three small molecules showed similar PAEs against the E. coli strains. Figure 2. View large Download slide PAEs observed when treating E. coli strains BW25113 (top), MC4100 (middle) and ATCC 25922 (bottom) (7.5 × 106 cells/mL; 1 h, 37 °C, 600 rpm) with PrAMPs and antibiotic comparators at concentrations of 0.1 ×, 0.4 ×, 1 × and 4 × MIC (white, light grey, dark grey and black bars, respectively). Experiments were done twice in triplicate on different days. Hash signs indicate that none of the six individual curves reached half-maximum absorbance of the control within 22 h, whereas dollar signs indicate that this threshold was not reached in one to five replicates. For details see Table S1. Figure 2. View large Download slide PAEs observed when treating E. coli strains BW25113 (top), MC4100 (middle) and ATCC 25922 (bottom) (7.5 × 106 cells/mL; 1 h, 37 °C, 600 rpm) with PrAMPs and antibiotic comparators at concentrations of 0.1 ×, 0.4 ×, 1 × and 4 × MIC (white, light grey, dark grey and black bars, respectively). Experiments were done twice in triplicate on different days. Hash signs indicate that none of the six individual curves reached half-maximum absorbance of the control within 22 h, whereas dollar signs indicate that this threshold was not reached in one to five replicates. For details see Table S1. The PAEs of PrAMPs and standard antibiotics against K. pneumoniae ATCC 10031 were consistently <3 h at 0.1 × MIC, with the highest value obtained for Api88 (2.6 ± 0.3 h) (Figure 3 and Table S1). Fourfold higher concentrations increased the PAEs of Api88 (7.7 ± 0.4 h), Api137 (4.2 ± 0.6 h) and A3-APO (4.4 ± 0.9 h) significantly, whereas the PAEs of all other compounds remained <4 h. Interestingly, at MIC the PAE increased for Api88 (18.8 ± 1.2 h), Api795 (13.4 ± 4.7 h), Bac7(1-60) (18.2 ± 3.3 h) and A3-APO (13.1 ± 2.7 h) to >12 h, but for all other peptides remained below 12 h, with drosocin Hyp5 showing the weakest PAE of 4.2 ± 0.5 h. The PAEs of the three small-molecule antibiotics were even lower (1.3–3.6 h at MIC). At high concentrations (4 × MIC) most PrAMPs exhibited strong PAEs against K. pneumoniae (≥18 h); only drosocin Hyp5 (10.3 ± 0.6 h), pyrrhocoricin (9.8 ± 0.5 h) and Chex1Arg 20 (6.6 ± 0.7 h) showed lower values. Note that at these high concentrations both aminoglycosides (gentamicin and kanamycin) exhibited as strong PAEs as the best PrAMPs (≥19.5 h), whereas the PAE of chloramphenicol remained at 3.5 ± 0.2 h. Figure 3. View large Download slide PAEs obtained by incubating K. pneumoniae ATCC 10031 (7.5 × 106 cells/mL; 1 h, 37 °C, 600 rpm) with PrAMPs and antibiotic comparators at concentrations of 0.1 ×, 0.4 ×, 1 × and 4 × MIC (white, light grey, dark grey and black bars, respectively). Experiments were done twice in triplicate on different days. The PAE was calculated as T − C, with C being the time the control needed to reach half-maximum absorbance and T the time necessary for the treated sample to reach the half-maximum absorbance of the control. Hash signs indicate that none of the six individual curves reached half-maximum absorbance of the control within 22 h, whereas dollar signs indicate that this threshold was not reached in one to five replicates. For details see Table S1. Figure 3. View large Download slide PAEs obtained by incubating K. pneumoniae ATCC 10031 (7.5 × 106 cells/mL; 1 h, 37 °C, 600 rpm) with PrAMPs and antibiotic comparators at concentrations of 0.1 ×, 0.4 ×, 1 × and 4 × MIC (white, light grey, dark grey and black bars, respectively). Experiments were done twice in triplicate on different days. The PAE was calculated as T − C, with C being the time the control needed to reach half-maximum absorbance and T the time necessary for the treated sample to reach the half-maximum absorbance of the control. Hash signs indicate that none of the six individual curves reached half-maximum absorbance of the control within 22 h, whereas dollar signs indicate that this threshold was not reached in one to five replicates. For details see Table S1. When P. aeruginosa ATCC 27853 was inhibited with small-molecule antibiotics at concentrations up to MIC, the PAEs remained <4 h (Figure 4 and Table S1). This was also true for PrAMPs at sub-MIC, but at MIC the PAEs of PrAMPs ranged from 4.3 ± 0.3 h (Api137) to 8.7 ± 1.2 h (A3-APO). At 4 × MIC the PAEs were all >10 h, with Onc112 showing the best PAE (16.8 ± 1.3 h). Drosocin Hyp5 and pyrrhocoricin were not tested as they were virtually inactive against P. aeruginosa. Gentamicin and kanamycin exhibited lower PAEs than the PrAMPs (7.8 ± 0.7 and 8.9 ± 1.1 h). Chloramphenicol showed the weakest PAE of all compounds (2.1 ± 0.7 h). Figure 4. View large Download slide PAEs obtained for P. aeruginosa ATCC 27853 (7.5 × 106 cells/mL; 1 h, 37 °C, 600 rpm) when incubated with PrAMPs and antibiotic comparators at concentrations of 0.1 ×, 0.4 ×, 1 × and 4 × MIC (white, light grey, dark grey and black bars, respectively). Experiments were done at least twice in triplicate on different days. The PAE was calculated as T − C with C being the time the control needed to reach half-maximum absorbance and T the time necessary for the treated sample to reach the half-maximum absorbance of the control. Dollar signs indicate that one to five individual curves did not reach the half-maximum absorbance of the control within 22 h. For details see Table S1. Figure 4. View large Download slide PAEs obtained for P. aeruginosa ATCC 27853 (7.5 × 106 cells/mL; 1 h, 37 °C, 600 rpm) when incubated with PrAMPs and antibiotic comparators at concentrations of 0.1 ×, 0.4 ×, 1 × and 4 × MIC (white, light grey, dark grey and black bars, respectively). Experiments were done at least twice in triplicate on different days. The PAE was calculated as T − C with C being the time the control needed to reach half-maximum absorbance and T the time necessary for the treated sample to reach the half-maximum absorbance of the control. Dollar signs indicate that one to five individual curves did not reach the half-maximum absorbance of the control within 22 h. For details see Table S1. PAE and viable cell count The viable cell counts of E. coli and K. pneumoniae usually decreased at MIC by at least 1 log unit within an incubation period of 1 h and even further for conditions providing long PAEs (Figures S3, S4 and S5). The viable count of the P. aeruginosa control was much lower (2.7 ± 1.1 × 105 cfu/mL) than the mean of all other strains (5.1 ± 2.2 × 106 cfu/mL). Chloramphenicol always showed bacteriostatic activities. Drosocin Hyp5 and pyrrhocoricin were also bacteriostatic except for E. coli ATCC 25922 (P. aeruginosa not tested). Interestingly, Chex1Arg20 was bacteriostatic solely against K. pneumoniae, for which it also exhibited the lowest PAE. Gentamicin and kanamycin were always bactericidal at 4 × MIC. Strong bactericidal activities were observed for the long PrAMPs Bac7(1-60) and A3-APO, which can be explained by their dual mode of action, i.e. intracellular inhibition at MIC and increasing membrane permeability at peptide concentrations above the MIC.9 Interestingly, the bactericidal effects obtained for PrAMPs against E. coli ATCC 25922 and P. aeruginosa exceeded the effects of aminoglycosides. Discussion There are several studies about PAEs, but standardized protocols are missing. Thus, the bacteria used, bacterial and antibiotic concentrations, exposure times and methods for drug removal vary in the literature, with only a few publications relying on comparable conditions (for a detailed discussion see the Supplementary Discussion available as Supplementary data at JAC Online).37,E. coli ATCC 25922 is commonly used in PAE studies and the MIC values of gentamicin (0.5 mg/L) and chloramphenicol (2 mg/L) measured here for this strain were comparable to previous studies.42–44 However, the PAEs of 8.0 ± 0.7 h (4 × MIC) and 2.6 ± 0.6 h (1 × MIC) obtained here for gentamicin lasted much longer than in previous reports, i.e. 1.1 h (4 × MIC) and 0.5 h (1 × MIC),44 but were close to the 2 and 4 h reported for gentamicin at 5 × MIC.42,47 The PAE of 1 h reported for bacteriostatic chloramphenicol (5 × MIC) was close to the 2.0 ± 0.5 h measured here at 4 × MIC.42 These differences are most likely related to different laboratory routines, especially the different cell densities. Our sensitive method, which enables the determination of PAEs in ∼20 h, relied on the cell density recommended for MIC assays, i.e. 7.5 × 105 cells/mL. In contrast, the literature relies on much higher cell densities of 2 × 107 to 2 × 108 cells/mL, which results also in higher MIC values when measured for these conditions.13,48 Importantly, the high-throughput method described here enabled determination of the PAE for four different concentrations of 10 PrAMPs and three antibiotic comparators with high reproducibility. Unexpectedly, the PAEs of several PrAMPs determined at 0.4 × MIC exceeded 4 h, i.e. Onc72 (≥5.6 h) and A3-APO (≥4.6 h) against E. coli MC4100 and BW25113, pyrrhocoricin (4.4 h) against E. coli ATCC 25922 and Api88 (7.7 h) and Api137 (4.2 h) even against K. pneumoniae ATCC 10031. The PAEs determined for PrAMPs at MIC lasted typically for >4 h, with the highest PAEs (≥10 h) obtained against at least one E. coli strain and K. pneumoniae ATCC 10031 for Api88, Api137, Bac7(1-60) and A3-APO and for Chex1Arg20 against E. coli BW25113 and MC4100. The effects seen for small-molecule antibiotics at MIC were considerably weaker. Only gentamicin exceeded 4 h against E. coli MC4100. As expected, the PAEs further increased for higher peptide concentrations and almost all substances showed values >10 h against all strains at 4 × MIC, which is partially related to the bactericidal effect, i.e. reduced viable cell numbers after the incubation period. The same was true for gentamicin and kanamycin, possessing PAEs comparable to PrAMPs, whereas the PAEs of chloramphenicol remained <4 h, indicating its bacteriostatic properties. As a statistical analysis considering all tested substances and bacterial strains would provide a very large data set, we compared the apidaecin and oncocin derivatives, which possess similar structures and properties, as one group with the three antibiotics (small molecules) as a second group. The apidaecin/oncocin group showed significantly higher PAEs than the group of antibiotics at 0.4 × and 1 × MIC against all strains, underlining their high therapeutic potential (Figure S6). Drosocin Hyp5 and pyrrhocoricin exhibited the lowest PAEs among the PrAMPs, comparable to gentamicin and kanamycin. The PAEs of the other PrAMPs increased in the following order: oncocins < apidaecins < Bac7(1-60) = A3-APO. The long-lasting PAEs of the two long PrAMPs can be explained by their additional effects on bacterial membranes.9,12 Interestingly, some PrAMPs reduced the viable cell counts after an incubation period of 1 h at 4 × MIC more than the bactericidal aminoglycosides. Considering the long PAEs of PrAMPs observed here at MIC and sub-MIC and the bactericidal activity indicated by decreasing viable cell counts, PrAMPs probably belong to group 2 of the pharmacodynamic classification scheme, i.e. antibiotics with concentration-dependent bactericidal action and prolonged PAEs, such as aminoglycosides, with Cmax as limiting factor.39 Although the strong PAEs were unexpected, a recent study on bacterial uptake perfectly correlates to this observation. Apidaecins and oncocins bind rapidly to E. coli and translocate into the cytoplasm within minutes, as indicated by labelled peptides using flow cytometry and confirmed by unlabelled peptides using liquid chromatography.11,48,49 The authors could show that internalized Api137 remains in the cells without leaking into fresh medium using conditions close to the PAE assay applied here.48 In contrast, Onc112 was detected in the fresh medium. However, the authors argued that Onc112 might be released mostly from non-viable cells. These uptake studies and the current PAE data can nicely explain the apparent discrepancy between high in vivo efficacy and relatively low plasma levels.29,34,35 Presumably, PrAMPs enter bacteria rapidly in the host and inhibit the 70S ribosomes, as indicated by the strong PAE, even at MIC and partially also at sub-MIC concentrations. Thus, bacterial growth is stopped, allowing the host immune system to eradicate the pathogens before they start to grow again. It should be noted that Laszlo Otvos’s group has provided compelling evidence that their designer peptide A3-APO stimulates the mouse immune system as a third mode of action.50 In this respect, the efficacy of A3-APO might be best described as an artificial host defence peptide that relies mostly on immunomodulation, as demonstrated in wound and systemic infection models for Propionibacterium acnes, Staphylococcus aureus and Proteus mirabilis.7,51,52 Most recently, they showed that pre-infection treatment with transdermally administered A3-APO protects mice against intraperitoneal infections with A. baumannii.53 Apidaecins and oncocins have not been studied in vivo for immunomodulatory effects yet. In vitro, neither type of PrAMP showed immunomodulatory effects on unstimulated and LPS-stimulated murine dendritic cells or macrophages.54 In contrast, reduction of LPS-induced TNF-α release of human macrophages and monocytes by apidaecins indicates a mild anti-inflammatory effect.55 Independently, Tavano et al.56 reported for apidaecin 1b a differential immunomodulation of human macrophages, dendritic cells and macrophages. Additionally, A3-APO increases the levels of anti-inflammatory cytokines IL-4 and IL-10 in human peripheral blood mononuclear cells, possibly indicating a reduction in inflammation at the site of an infection.52 The PAEs presented here provide an additional hypothesis besides immunomodulatory effects that can explain the good in vivo efficacies of PrAMPs, which are not wholly expected, in different infection models despite the fast clearance rates reported for apidaecins, oncocins, A3-APO and Chex1Arg20 in mice. This highlights again that MIC values determined for AMPs in vitro cannot be simply used to predict in vivo efficacies, as often assumed in the literature. Instead MIC values should be seen as one important criterion among other parameters to be considered. Acknowledgements We thank Tina Goldbach and Dr Martina Bluhm for technical support in peptide synthesis and the Federal Ministry of Education and Research (BMBF), the European Fund for Regional Structure Development (EFRE) and the German Research Foundation (DFG) for financial support. Funding This work was supported by the Federal Ministry of Education and Research (BMBF; grant number 01GU1104A), the European Fund for Regional Structure Development (EFRE; European Union and Free State of Saxony; grant numbers 100105139 and 100127675) and the German Research Foundation (DFG; grant number INST 268/289-1). Transparency declarations D. K. was a part-time co-worker of AMP Therapeutics (Leipzig, Germany). R. H. is a co-founder of AMP Therapeutics and a member of its scientific advisory board. L. H.: none to declare. Author contributions L. H. conducted the experiments. L. H. and D. K. designed the experiments. L. H., D. K. and R. H. wrote the manuscript. Supplementary data Figures S1 to S6, Table S1 and Supplementary Discussion are available as Supplementary data at JAC Online. References 1 Falagas ME, Mavroudis AD, Vardakas KZ. The antibiotic pipeline for multi-drug resistant Gram-negative bacteria: what can we expect? Expert Rev Anti Infect Ther  2016; 14: 747– 63. Google Scholar CrossRef Search ADS   2 Casteels P, Ampe C, Jacobs F et al.   Apidaecins: antibacterial peptides from honeybees. EMBO J  1989; 8: 2387– 91. 3 Knappe D, Cassone M, Nollmann FI et al.   Hydroxyproline substitutions stabilize non-glycosylated drosocin against serum proteases without challenging its antibacterial activity. Protein Pept Lett  2014; 21: 321– 9. Google Scholar CrossRef Search ADS   4 Knappe D, Piantavigna S, Hansen A et al.   Oncocin (VDKPPYLPRPRPPRRIYNR-NH2): a novel antibacterial peptide optimized against Gram-negative human pathogens. J Med Chem  2010; 53: 5240– 7. Google Scholar CrossRef Search ADS   5 Giacometti A, Cirioni O, Del Prete MS et al.   In vitro effect on Cryptosporidium parvum of short-term exposure to cathelicidin peptides. J Antimicrob Chemother  2003; 51: 843– 7. Google Scholar CrossRef Search ADS   6 Otvos LJr, Wade JD, Lin F et al.   Designer antibacterial peptides kill fluoroquinolone-resistant clinical isolates. J Med Chem  2005; 48: 5349– 59. Google Scholar CrossRef Search ADS   7 Ostorhazi E, Voros E, Nemes-Nikodem E et al.   Rapid systemic and local treatments with the antibacterial peptide dimer A3-APO and its monomeric metabolite eliminate bacteria and reduce inflammation in intradermal lesions infected with Propionibacterium acnes and meticillin-resistant Staphylococcus aureus. Int J Antimicrob Agents  2013; 42: 537– 43. Google Scholar CrossRef Search ADS   8 Scocchi M, Mardirossian M, Runti G et al.   Non-membrane permeabilizing modes of action of antimicrobial peptides on bacteria. Curr Top Med Chem  2016; 16: 76– 88. Google Scholar CrossRef Search ADS   9 Podda E, Benincasa M, Pacor S et al.   Dual mode of action of Bac7, a proline-rich antibacterial peptide. Biochim Biophys Acta  2006; 1760: 1732– 40. Google Scholar CrossRef Search ADS   10 Mattiuzzo M, Bandiera A, Gennaro R et al.   Role of the Escherichia coli SbmA in the antimicrobial activity of proline-rich peptides. Mol Microbiol  2007; 66: 151– 63. Google Scholar CrossRef Search ADS   11 Krizsan A, Knappe D, Hoffmann R. Influence of the yjiL-mdtM gene cluster on the antibacterial activity of proline-rich antimicrobial peptides overcoming Escherichia coli resistance induced by the missing SbmA transporter system. Antimicrob Agents Chemother  2015; 59: 5992– 8. Google Scholar CrossRef Search ADS   12 Graf M, Mardirossian M, Nguyen F et al.   Proline-rich antimicrobial peptides targeting protein synthesis. Nat Prod Rep  2017; 34: 702– 11. Google Scholar CrossRef Search ADS   13 Bluhm ME, Schneider VA, Schafer I et al.   N-terminal Ile-Orn- and Trp-Orn-motif repeats enhance membrane interaction and increase the antimicrobial activity of apidaecins against Pseudomonas aeruginosa. Front Cell Dev Biol  2016; 4: 39. Google Scholar CrossRef Search ADS   14 Knappe D, Ruden S, Langanke S et al.   Optimization of oncocin for antibacterial activity using a SPOT synthesis approach: extending the pathogen spectrum to Staphylococcus aureus. Amino Acids  2016; 48: 269– 80. Google Scholar CrossRef Search ADS   15 Otvos LJr. The short proline-rich antibacterial peptide family. Cell Mol Life Sci  2002; 59: 1138– 50. Google Scholar CrossRef Search ADS   16 Czihal P, Knappe D, Fritsche S et al.   Api88 is a novel antibacterial designer peptide to treat systemic infections with multidrug-resistant Gram-negative pathogens. ACS Chem Biol  2012; 7: 1281– 91. Google Scholar CrossRef Search ADS   17 Knappe D, Kabankov N, Hoffmann R. Bactericidal oncocin derivatives with superior serum stabilities. Int J Antimicrob Agents  2011; 37: 166– 70. Google Scholar CrossRef Search ADS   18 Volke D, Krizsan A, Berthold N et al.   Identification of Api88 binding partners in Escherichia coli using a photoaffinity-cross-link strategy and label-free quantification. J Proteome Res  2015; 14: 3274– 83. Google Scholar CrossRef Search ADS   19 Krizsan A, Volke D, Weinert S et al.   Insect-derived proline-rich antimicrobial peptides kill bacteria by inhibiting bacterial protein translation at the 70 S ribosome. Angew Chem Int Ed Engl  2014; 53: 12236– 9. Google Scholar CrossRef Search ADS   20 Krizsan A, Prahl C, Goldbach T et al.   Short proline-rich antimicrobial peptides inhibit either the bacterial 70S ribosome or the assembly of its large 50S subunit. Chembiochem  2015; 16: 2304– 8. Google Scholar CrossRef Search ADS   21 Florin T, Maracci C, Graf M et al.   An antimicrobial peptide that inhibits translation by trapping release factors on the ribosome. Nat Struct Mol Biol  2017; 24: 752– 7. Google Scholar CrossRef Search ADS   22 Gagnon MG, Roy RN, Lomakin IB et al.   Structures of proline-rich peptides bound to the ribosome reveal a common mechanism of protein synthesis inhibition. Nucleic Acids Res  2016; 44: 2439– 50. Google Scholar CrossRef Search ADS   23 Roy RN, Lomakin IB, Gagnon MG et al.   The mechanism of inhibition of protein synthesis by the proline-rich peptide oncocin. Nat Struct Mol Biol  2015; 22: 466– 9. Google Scholar CrossRef Search ADS   24 Seefeldt AC, Nguyen F, Antunes S et al.   The proline-rich antimicrobial peptide Onc112 inhibits translation by blocking and destabilizing the initiation complex. Nat Struct Mol Biol  2015; 22: 470– 5. Google Scholar CrossRef Search ADS   25 Seefeldt AC, Graf M, Perebaskine N et al.   Structure of the mammalian antimicrobial peptide Bac7(1-16) bound within the exit tunnel of a bacterial ribosome. Nucleic Acids Res  2016; 44: 2429– 38. Google Scholar CrossRef Search ADS   26 Knappe D, Fritsche S, Alber G et al.   Oncocin derivative Onc72 is highly active against Escherichia coli in a systemic septicaemia infection mouse model. J Antimicrob Chemother  2012; 67: 2445– 51. Google Scholar CrossRef Search ADS   27 Szabo D, Ostorhazi E, Binas A et al.   The designer proline-rich antibacterial peptide A3-APO is effective against systemic Escherichia coli infections in different mouse models. Int J Antimicrob Agents  2010; 35: 357– 61. Google Scholar CrossRef Search ADS   28 Hoffmann R, Bulet P, Urge L et al.   Range of activity and metabolic stability of synthetic antibacterial glycopeptides from insects. Biochim Biophys Acta  1999; 1426: 459– 67. Google Scholar CrossRef Search ADS   29 Ostorhazi E, Nemes-Nikodem E, Knappe D et al.   In vivo activity of optimized apidaecin and oncocin peptides against a multiresistant, KPC-producing Klebsiella pneumoniae strain. Protein Pept Lett  2014; 21: 368– 73. Google Scholar CrossRef Search ADS   30 Benincasa M, Pelillo C, Zorzet S et al.   The proline-rich peptide Bac7(1-35) reduces mortality from Salmonella typhimurium in a mouse model of infection. BMC Microbiol  2010; 10: 178. Google Scholar CrossRef Search ADS   31 Ostorhazi E, Rozgonyi F, Sztodola A et al.   Preclinical advantages of intramuscularly administered peptide A3-APO over existing therapies in Acinetobacter baumannii wound infections. J Antimicrob Chemother  2010; 65: 2416– 22. Google Scholar CrossRef Search ADS   32 Ostorhazi E, Rozgonyi F, Szabo D et al.   Intramuscularly administered peptide A3-APO is effective against carbapenem-resistant Acinetobacter baumannii in mouse models of systemic infections. Biopolymers  2011; 96: 126– 9. Google Scholar CrossRef Search ADS   33 Knappe D, Adermann K, Hoffmann R. Oncocin Onc72 is efficacious against antibiotic-susceptible Klebsiella pneumoniae ATCC 43816 in a murine thigh infection model. Biopolymers  2015; 104: 707– 11. Google Scholar CrossRef Search ADS   34 Schmidt R, Ostorhazi E, Wende E et al.   Pharmacokinetics and in vivo efficacy of optimized oncocin derivatives. J Antimicrob Chemother  2016; 71: 1003– 11. Google Scholar CrossRef Search ADS   35 Schmidt R, Knappe D, Wende E et al.   In vivo efficacy and pharmacokinetics of optimized apidaecin analogs. Front Chem  2017; 5: 15. Google Scholar CrossRef Search ADS   36 Craig WA. The postantibiotic effect. Clin Microbiol Newslett  1991; 13: 121– 4. Google Scholar CrossRef Search ADS   37 MacKenzie FM, Gould IM. The post-antibiotic effect. J Antimicrob Chemother  1993; 32: 519– 37. Google Scholar CrossRef Search ADS   38 Lowdin E, Odenholt-Tornqvist I, Bengtsson S et al.   A new method to determine postantibiotic effect and effects of subinhibitory antibiotic concentrations. Antimicrob Agents Chemother  1993; 37: 2200– 5. Google Scholar CrossRef Search ADS   39 Levison ME, Levison JH. Pharmacokinetics and pharmacodynamics of antibacterial agents. Infect Dis Clin North Am  2009; 23: 791– 815, vii. Google Scholar CrossRef Search ADS   40 Craig W. Pharmacodynamics of antimicrobial agents as a basis for determining dosage regimens. Eur J Clin Microbiol Infect Dis  1993; 12 Suppl 1: S6– 8. Google Scholar CrossRef Search ADS   41 Tenson T, Mankin A. Antibiotics and the ribosome. Mol Microbiol  2006; 59: 1664– 77. Google Scholar CrossRef Search ADS   42 Stubbings WJ, Bostock JM, Ingham E et al.   Assessment of a microplate method for determining the post-antibiotic effect in Staphylococcus aureus and Escherichia coli. J Antimicrob Chemother  2004; 54: 139– 43. Google Scholar CrossRef Search ADS   43 Zhanel GG, Karlowsky JA, Davidson RJ et al.   Effect of pooled human cerebrospinal fluid on the postantibiotic effects of cefotaxime, ciprofloxacin, and gentamicin against Escherichia coli. Antimicrob Agents Chemother  1992; 36: 1136– 9. Google Scholar CrossRef Search ADS   44 Dominguez MC, de La Rosa M, Borobio MV. Application of a spectrophotometric method for the determination of post-antibiotic effect and comparison with viable counts in agar. J Antimicrob Chemother  2001; 47: 391– 8. Google Scholar CrossRef Search ADS   45 Baba T, Ara T, Hasegawa M et al.   Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the Keio collection. Mol Syst Biol  2006; 2: 2006.0008. Google Scholar CrossRef Search ADS   46 Fass RJ, Barnishan J. Minimal inhibitory concentrations of 34 antimicrobial agents for control strains Escherichia coli ATCC 25922 and Pseudomonas aeruginosa ATCC 27853. Antimicrob Agents Chemother  1979; 16: 622– 4. Google Scholar CrossRef Search ADS   47 Stubbings W, Bostock J, Ingham E et al.   Mechanisms of the post-antibiotic effects induced by rifampicin and gentamicin in Escherichia coli. J Antimicrob Chemother  2006; 58: 444– 8. Google Scholar CrossRef Search ADS   48 Holfeld L, Hoffmann R, Knappe D. Correlating uptake and activity of proline-rich antimicrobial peptides in Escherichia coli. Anal Bioanal Chem  2017; 409: 5581– 92. Google Scholar CrossRef Search ADS   49 Berthold N, Hoffmann R. Cellular uptake of apidaecin 1b and related analogs in Gram-negative bacteria reveals novel antibacterial mechanism for proline-rich antimicrobial peptides. Protein Pept Lett  2014; 21: 391– 8. Google Scholar CrossRef Search ADS   50 Otvos L. Racing on the wrong track. Front Chem  2017; 5: 42. Google Scholar CrossRef Search ADS   51 Otvos LJr, Ostorhazi E. Therapeutic utility of antibacterial peptides in wound healing. Expert Rev Anti Infect Ther  2015; 13: 871– 81. Google Scholar CrossRef Search ADS   52 Ostorhazi E, Holub MC, Rozgonyi F et al.   Broad-spectrum antimicrobial efficacy of peptide A3-APO in mouse models of multidrug-resistant wound and lung infections cannot be explained by in vitro activity against the pathogens involved. Int J Antimicrob Agents  2011; 37: 480– 4. Google Scholar CrossRef Search ADS   53 Ostorhazi E, Horvath A, Szabo D et al.   Transdermally administered proline-arginine-rich host defense peptides show systemic efficacy in a lethal mouse bacteremia model. Amino Acids  2017; 49: 1647– 51. Google Scholar CrossRef Search ADS   54 Fritsche S, Knappe D, Berthold N et al.   Absence of in vitro innate immunomodulation by insect-derived short proline-rich antimicrobial peptides points to direct antibacterial action in vivo. J Pept Sci  2012; 18: 599– 608. Google Scholar CrossRef Search ADS   55 Keitel U, Schilling E, Knappe D et al.   Effect of antimicrobial peptides from Apis mellifera hemolymph and its optimized version Api88 on biological activities of human monocytes and mast cells. Innate Immun  2013; 19: 355– 67. Google Scholar CrossRef Search ADS   56 Tavano R, Segat D, Gobbo M et al.   The honeybee antimicrobial peptide apidaecin differentially immunomodulates human macrophages, monocytes and dendritic cells. J Innate Immun  2011; 3: 614– 22. Google Scholar CrossRef Search ADS   © The Author 2017. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please email: journals.permissions@oup.com.

Journal

Journal of Antimicrobial ChemotherapyOxford University Press

Published: Apr 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Unlimited reading

Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere.

Stay up to date

Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates.

Organize your research

It’s easy to organize your research with our built-in tools.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve Freelancer

DeepDyve Pro

Price
FREE
$49/month

$360/year
Save searches from
Google Scholar,
PubMed
Create lists to
organize your research
Export lists, citations
Read DeepDyve articles
Abstract access only
Unlimited access to over
18 million full-text articles
Print
20 pages/month
PDF Discount
20% off