Modified poly(L-lysine)-based structures as novel antimicrobials for diabetic foot infections, an in-vitro study

Ethics statement

Ethical approval was not sought as it is not required for research on micro-organisms, which are not considered as belonging to the patient or any individual person. No human tissue was used in this study and no additional patient specimen processing or investigations were undertaken. No patient data was collected, processed or stored. Swabs were taken in 2019 at Beaumont Hospital, Dublin, with the patient’s verbal consent, by healthcare staff, from the wound or base of foot ulcers (for DFW) for clinical investigation, as part of their routine clinical care to support the diagnosis of wound infection/diabetic foot infection. Swabs are subsequently routinely processed for recovery and identification of colonizing/infecting organisms by the Beaumont Hospital Microbiology Laboratory. When either S. aureus or P. aeruginosa were identified, pure cultures of these bacteria were provided to the researcher. The study bacteria were received entirely anonymously without any associated patient identifier or data.

PLL-polymer synthesis

This research used the following:

M03079 - H-Lys(Z)-OH - ε-carbobenzyloxy-L-lysine

M02961 - H-Phe-OH - L-phenylalanine

M03002 - H-lle-OH - L-isoleucine

M03100 - H-Tyr(Bzl)-OH - O-benzyl-L-tyrosine

All chemicals and solvents were obtained from Sigma Aldrich (Ireland) unless otherwise noted. ε-carbobenzyloxy-L-lysine (M03079), L-isoleucine (M03002), L-phenylalanine (M02961) and O-benzyl-L-tyrosine (M03100) were purchased from Fluorochem (UK)

The polypeptides and copolypeptides were synthesised via N-carboxyanhydride ring-opening polymerisation (NCA ROP) as previously described¹⁴. For synthesis of ε-carbobenzyloxy-L-lysine (ZLL) NCA: ε-carbobenzyloxy-L-lysine (15 g, 53.51 mmol) and α-pinene (147524, 18.22 g, 133.78 mmol) were suspended in 180 mL of dry tetrahydrofuran (THF) and heated under reflux. A solution of triphosgene (330752, 7.15 g, 24.08 mmol) in 30 mL dry THF was added drop-wise to the suspension. The suspension was refluxed until all solids disappeared and the solution became clear (5 hours). The solution was then cooled, filtered and 2/3 of the volume was removed under vacuum. It was then precipitated by addition of 250 mL hexane and stored overnight at -18 °C. The NCA solid was dried, redissolved in dry ethyl acetate and filtered. The NCA solution was then recrystallised thrice in ethyl acetate/hexane (1:1.5) and subsequently washed with hexane to remove any trace impurities. It was vacuum dried to afford a colourless fluffy solid (yield 83%). The NCAs of L-isoleucine (LI), L-phenylalanine (LP) and O-benzyl-L-tyrosine (BLT) were synthesised using the same method. For synthesis of 8arm-star-poly(L-lysine) (G2(8)PLL₂₀) – (representative procedure): ε-carbobenzyloxy-L-lysine (ZLL) NCA (1 g, 2.27 mmol) was dissolved in 15 mL CHCl₃ and 3 mL dimethylformamide. G2 poly(propylene imine) (PPI) dendrimer (15.87 mg, 2.04 × 10^-2 mmol) was dissolved in 3 mL of CHCl₃ and it was then added quickly to the NCA solution. The solution was allowed to stir until Fourier-transform infrared spectroscopy analysis confirmed total consumption of the ZLL NCA monomer. The polymer was then precipitated into excess diethyl ether and dried under vacuum in a dessicator (yield 85%). To deprotect the polypeptide, 500 mg of G2(8)PLL₂₀ was dissolved in 10 mL of trifluoroacetic acid (TFA), followed by dropwise addition of 3 mL of 33 wt% hydrobromic acid (HBr) in acetic acid. The solution was stirred for 18 hours followed by precipitation into excess diethyl ether (150 mL). After drying in vacuo, it was dissolved in deionised water and loaded into a 3500 molecular weight cut off (MWCO) membrane (Thermofisher Scientific, Ireland). Dialysis was performed against deionised water for four days with frequent water replacement, followed by lyophilisation (yield 72%). The other polypepft-irtides were prepared in the same way, using different PPI dendrimers to form G3 and G4 PLL derivatives respectively. The addition of NCA comonomers (LI NCA, LP NCA or BLT NCA) along with ZLL NCA using the same method afforded the statistical copolypeptides containing hydrophobic amino acid repeat units. In this way, three star PLLs were generated: G2(8)PLL₂₀ (8-star-PLL) (G2 = generation 2 poly(propylene imine) dendrimer (PPI) core, (8) = 8 polymeric L-lysine arms with 20 L-lysine subunits per arm, M_n = 21,400 g/mol); G3(16) PLL₁₀ (16-star-PLL) (G3 = generation 3 PPI core, (16) = 16 polymeric L-lysine arms with 10 L-lysine subunits per arm, M_n = 22,300 g/mol); and G4(32)PLL₅ (32-star-PLL) (G4 = generation 4 PPI core, (32) = 32 polymeric L-lysine arms with 5 L-lysine subunits per arm, M_n= 24,200 g/mol). The equivalent linear polymer PLL₁₆₀ was a comparator for this series (no dendrimer core) M_n = 20,600 g/mol. A further G3(16)PLL₂₀ series was generated with an additional five hydrophobic amino acid residues attached to the star arms as follows: G3(16)PLL₂₀-co-PLI₅ (G3=generation 3 PPI core, 16 = polymeric L-lysine arms, PLL20 = 20 poly(L-lysine) per arm, co-PLI₅ = 5 poly (L-isoleucine) per arm, M_n = 50,300 g/mol, G3(16)PLL₂₀-co-PLP₅, co-PLP₅ = 5 poly (L-phenylalanine) per arm, M_n = 53,000 g/mol and G3(16)PLL₂₀-co-PLT_5, co-PLT₅= 5 poly (L-tyrosine) units per arm, M_n = 54,300 g/mol.

Bacterial strains and isolates

A laboratory strain from the American Type Culture Collection, S. aureus ATCC 25923, and the widely used P. aeruginosa strain PAO1 (ATCC 156920) which originates from a wound infection¹⁵ were used to establish assay conditions for antimicrobial evaluation of polymers. Two collections of clinical isolates were further evaluated. One consisted of S. aureus and P. aeruginosa recovered from swabs taken as part of routine clinical care from patients with suspected infection during monthly multidisciplinary diabetic foot clinics taking place in Beaumont Hospital, Dublin in 2019 for microbiological diagnosis. The second consisted of 10 S. aureus, five of which were MRSA, which were from confirmed wound infections from the same hospital in the same collection period (2019). All swabs were processed by the Beaumont Hospital Microbiology Laboratory, Dublin and were provided anonymized, as pure cultures. Organism identity was confirmed by matrix-assisted laser desorption/ionization- time-of-flight mass spectrometry (MALDI-TOF) using a MALDI Biotyper (Brüker). MRSA was confirmed by the growth of pink colonies on MRSASelect^TM agar (BioRad, UK). In total 30 S. aureus (of which 10 were MRSA) and five P. aeruginosa were collected. For testing novel antimicrobials, clinical isolates were selected randomly from those available to the researchers with numbers chosen as a cross-section, representative of those colonizing and infecting DFIs. The lower numbers of P. aeruginosa available compared to S. aureus recovered are also representative of the trends in recovered pathogens. Within the isolate collection, for determination of statistical significance, e.g. comparison of the activity of linear with star PLLs, biofilm formation, greater than three clinical isolates in each group was considered appropriate.

Bactericidal assay

Assays were performed as described by Forde et al.¹⁶ with modifications. Briefly, isolated colonies from overnight growth of S. aureus on Columbia Blood Agar (CBA) or P. aeruginosa on MacConkey agar were suspended in PBS to the density of 0.5 MacFarland Standard using a Densichek Meter (approximately 5 × 10⁷ colony forming units (CFU)/ml). Suspensions were further diluted in the assay buffer (10mM potassium phosphate buffer, pH7.4, containing 0.2% w/v bovine serum albumin (BSA)). Assays contained approximately 10⁵ CFU/ml bacteria and varying concentrations of polymers (0.001μM-10μM) in a total volume of 100μl of assay buffer. Following incubation at 37°C for 1 hour, 900 μl of NaCl (0.95% w/v) was added to assays and 100 μl aliquots were spread onto Mueller Hinton Agar. Plates were incubated at 37°C overnight in a static incubator and resulting colonies were counted. Where numbers were ‘too numerous to count’, further dilutions were plated. Killing activity was determined based on CFU/ml from assays containing PLL polymers (log¹⁰ reduction) compared to CFU/ml on control plates containing no polymer. Rifampicin or gentamicin were used as comparators where appropriate against S. aureus and P. aeruginosa, respectively.

Biofilm viability assay

Effects of the polymers on biofilms were investigated as described previously¹⁷. Colonies from overnight cultures were isolated and used to prepare suspensions to the density of a 0.5 McFarland standard using a Densichek meter (bioMérieux, Ireland). The suspensions were further diluted 1/100 in tryptic soy broth (TSB) containing 0.5% glucose and 100 μL added to 96 well polystyrene plates. Plates were incubated for 24 hours at 37°C. Following washing (3 times with distilled water), the biofilms were treated with various PLL polymers (50 μM) by the addition of 100 μL to wells. Rifampicin or gentamicin (100 μM) was added to biofilm wells as comparators and 100 μl of KPB, pH 7.4 was added to biofilm wells as controls. Wells containing TSB only were included as media controls. The treated plates were incubated at 37°C for 30 minutes. For biofilm viability measurement after treatment, a stock solution of the non-fluorescent redox dye resazurin at 440 μM was diluted 1/5 in TSB and 100 µL was added to each well. Plates were incubated at 37°C for 1 hour in the dark. Fluorescence at 544 nm excitation and 590 nm emission was measured using a Perkin Elmer 2030 Multi-label Reader Viktor X3 spectrophotometer.

Star and linear polypeptide structures can be synthesised

A set of star-shaped poly(L-lysine) structures with 8, 16 and 32 arms were synthesised by the ring-opening polymerisation of ε-carbobenzyloxy-L(lysine) N-carboxyanhydride (ZLL NCA), followed by subsequent deprotection and purification^18,19. The total number of PLL units was kept constant at 160 for all star polypeptides resulting in decreasing arm length with increasing number of arms (Figure 1). A linear PLL was also obtained. In addition, 16-arm statistical copolypeptides comprising 20 repeating units of PLL and 5 repeating units of different hydrophobic amino acids were synthesised. The latter were selected from isoleucine, phenylalanine and tyrosine. These and similar structures were previously investigated extensively as delivery tools for applications in tissue engineering where their versatility in delivering therapeutic cargos into mesenchymal stem cells (MSC) has been demonstrated¹⁴.

Figure 1. Structures of star poly(L-lysine) homo- and copolypeptides.

Linear PLL polymer has similar antimicrobial activity to equivalent PLLs with star-shaped architecture

Potent bactericidal activity was demonstrated against the reference S. aureus strain ATCC 25923 and P. aeruginosa strain PAO1. Preliminary testing of 10 μM PLLs, linear (PLL₁₆₀) Vs star architecture (G2(8)PLL₂₀) for 1 hour against these bacteria resulted in killing greater than the upper limit of the assay (log 5) compared to log 2 reduction in CFU/ml using the antibiotics rifampicin or gentamicin (Figure 2). A lower concentration of 1 μM was therefore investigated to observe differential activity among PLLs. Incubation with PLL-polymers at 1 μM for 1 hour resulted in a 2-4 log reduction in CFU/ml for S. aureus with the greatest activity for the linear PLL₁₆₀ (4.23 log) and the least for the star architecture with 8 arms around the central core (G2(8)PLL₂₀ = 2.29 log). Improved activity was found as the arm numbers increased to 16 and 32. At 1 μM, 2-3 log reduction in CFU/ml was found for PAO1 with little difference in activity with linear or any of the star architectures, in terms of number of star arms (8, 16, 32) or length of star arms (Figure 3).

Figure 2. Bactericidal activity of poly-L-lysine polymers against S. aureus and P. aeruginosa laboratory strains compared to antibiotics.

Log CFU/ml for a fixed concentration of bacteria following incubation with 10μM PLL polymers in 10mM potassium phosphate buffer (1 hour, 37^oC), 0.2% BSA. S. aureus ATCC 25923 (light grey bars), P. aeruginosa PAO1 (dark grey bars). Rifampicin (RIF) and gentamicin (GENT) were used at 10 μM for comparison. Data shown are the mean ± SEM of three assays carried out in duplicate.

Figure 3. Bactericidal activity of poly-L-lysine polymers against S. aureus and P. aeruginosa laboratory strains.

Log reduction in CFU/ml for a fixed concentration of bacteria following incubation with 1μM PLL polymers in 10mM potassium phosphate buffer (1 hour, 37^oC), 0.2% BSA. S. aureus ATCC 25923 (light grey bars), P. aeruginosa PAO1 (dark grey bars). Data shown are the mean ± SEM of three assays carried out in duplicate.

Antimicrobial activity is retained on addition of hydrophobic amino acids to the star architecture

Using the G3(16)PLL₂₀ series as a model, further modification of the star arms by the addition of copolypeptides of the amino acids tyrosine, isoleucine or phenylalanine to the poly-lysine arms were made. These modifications did not significantly change the antimicrobial activity against S. aureus (p value ≥ 0.25) which remained at 3-4 log following 1 hour incubation with 1 μM (Figure 4).

Figure 4. Comparative bactericidal activity of poly-L-lysine G3(16) copolymers series with hydrophobic amino acid isoleucine, tyrosine and phenylalanine.

Log reduction in CFU/ml with respect to control for a fixed concentration of S. aureus ATCC 25923, following incubation with 1μM polymer/copolymer in 10mM potassium phosphate buffer (1 hour, 37^oC), 0.2% BSA. Data shown are the mean ± SEM of three assays carried out in duplicate.

Clinical isolates from wound infections and DFW are susceptible to poly(L-lysine) polymers

Clinical isolates recovered from colonized wounds and infections often have different behaviours and properties to laboratory strains. Using PLL₁₆₀ and G2(8)PLL₂₀ again at 1 μM to reveal potential potency differentials among isolates, both linear and star PLLs remained active against S. aureus (n=10 isolates including 5 MRSA) recovered from confirmed wound infections with mean values of 2.3 and 2.7 log reduction in CFU/ml respectively (Figure 5). As the linear PLL-polymer was the most active at this lower concentration, it was selected for investigation across a further collection of clinical isolates from DFW/DFI. Potent bactericidal activity was observed for S. aureus and P. aeruginosa clinical isolates recovered from DFW/DFI with some variability between isolates (Figure 6). For S. aureus, the mean log reduction was 3.04 ± 0.16 (range 1.86 to 4.22, n=20), for P. aeruginosa isolates the mean log reduction was 3.96 ± 0.82 (range 2.35 to 4.29, n=5).

Figure 5. Comparison of Bactericidal activity of poly-L-lysine polymers, PLL₁₆₀ and G2(8)PLL₂₀ against S. aureus clinical isolates from wound infections.

Log reduction in CFU/ml for a fixed concentration of bacteria following incubation with 1μM PLL or G2(8)PLL₂₀ polymer in 10mM potassium phosphate buffer (1 hour, 37^oC), 0.2% BSA compared to control (no treatment). Assays carried out in duplicate on three occasions for n=10 S. aureus clinical isolates. NS = not statistically significant, p > 0.1 by unpaired t-test.

Figure 6. Bactericidal activity of linear PLL₁₆₀ against S. aureus and P. aeruginosa isolates from suspected diabetic foot infections.

Log reduction in CFU/ml for a fixed concentration of bacteria following incubation with 1 μM PLL₁₆₀ in 10mM potassium phosphate buffer (1 hour, 37^oC), 0.2% BSA compared to control (no treatment). Data shown are the mean ± SEM for assays carried out in duplicate on three occasions per isolate. Hatched bars indicate MRSA isolates.

Anti-biofilm activity of PLL-polymers

The G3(16)PLL₂₀ series, including copolymer modifications of the star arms showed statistically significant reduction in biofilm viability over a short time of exposure (30 minutes) to biofilms of S. aureus clinical isolates (p ≤ 0.05 (linear and star G3(16)PLL₂₀ or p ≤ 0.01, co-polymer series). However, only one candidate, G3(16)PLL₂₀-co-PLI₅ had anti-biofilm activity against P. aeruginosa isolates (Figure 7). Rifampicin and gentamicin in comparison demonstrated negligible anti-biofilm activity in-vitro under these conditions.

Figure 7. Investigation of PLL polymer-induced loss of biofilm viability by resazurin staining of 24 h biofilms.

S. aureus reference strain ATCC 25923 (a) and S. aureus DFW/DFI clinical isolates (n=10) (b) or P. aeruginosa DFW/DFI clinical isolates (n=5) recovered from diabetic foot infections. Biofilms were treated with each polymer (50 μM) for 30 minutes at 37^oC using at least six replicates per assay. Data shown are the mean ± SEM. Rifampicin/gentamicin concentration = 100 μM, * p ≤ 0.05, ** p ≤ 0.01.

Underlying data

Zenodo: Modified poly(L-lysine)-based structures as novel antimicrobials for diabetic foot infections, an in-vitro study. https://doi.org/10.5281/zenodo.5537160

This project contains the following files:

Data are available under the terms of the Creative Commons Attribution 4.0 International license (CC-BY 4.0).