The synthesis of polymeric dual-functional antimicrobial surface based on poly(2-methyl-2-oxazoline)

There is a high interest in the development of antimicrobial coatings to ﬁ ght bacterial infections. We present the development of dual-functional antimicrobial surface, in which a biopassive platform was functionalized with bioactive compounds on the surface, using a graft copolymer system poly(L-lysine)-graft-poly(2-methyl-2-oxazoline)-quarternery ammonium compound (PLL-g-PMOXA-QAC). Alkyne functionality was introduced to the PMOXA chain at α -terminus by initiating the living cationic polymerization of 2-methyl-2-oxazoline with a propargylic-initiator. The reaction was terminated with carboxy derivative-terminator that allows grafting of the polymeric chain from the β -terminus to poly(L-lysine) (PLL) backbone, resulting in graft copolymer alkynyl PLL-g-PMOXA. The conjugation between alkynyl PLL-g-PMOXA and QAC was then performed using click reaction. The chemical structures of the polymers were characterized by MALDI-TOF spectrometry and NMR spectroscopy. The results demonstrate that we have successfully synthesized PLL-g-PMOXA-QAC copolymer with grafting density (number of lysine/number of PMOXA) of 0.33. The resulting PLL-g-PMOXA-QAC copolymer was then immobilized onto carboxylated tissue cultured polystyrene (TCPS) surface and exposed to bacteria solution to test its dual-functional properties. Preliminary live-and-dead bacteria study indicates dual-functionality of the PLL-g-PMOXA-QAC-coated surface.


Introduction
The development of a biofilm on a medical device and the inherent diffi culties of its suppression have wide-ranging consequences for the patients. Among them, severe pain, need for explantation, and requirement for a revision surgery are the most common ones (Schneider et al., 2008). Intra-vascular catheter-related infections are also a major problem and cause of morbidity and mortality. The annual infection rate for cardiovascular implants is even higher (7.4%). Bacterial contamination is also a widespread problem in the population using devices such as contact lenses, dental implant materials, intrauterine contraceptive devices and vaginal tampons. It is also a problem related to the safety of tools and materials used in medicine or surgery. There is therefore a great interest in approaches to prevent or reduce the risk of infections related to implanted materials.
Currently, several approaches have been developed to protect biomaterials from bacterial contamination (Xie et al., 2011). Based on the type of interaction with bacteria, one may distinguish between 'biopassive' And 'bioactive' biomaterial surfaces. Biopassive surfaces prevent the adhesion of bacteria but do not actively interfere with them. In contrast, bioactive surfaces aim at killing bacteria upon contact or at interfering with the quorum sensing between bacteria preventing further proliferation and maturation of biofi lms. A further distinction can be made between stably surface-immobilized antimicrobials and active release systems. Appropriate choice of systems depends on the type of antimicrobials and their mechanism of action. Regulatory aspects are also to be considered, in particular in regard to whether approval for a medical device of drug is required.
Materials having biopassive (nonfouling) properties prevent the adhesion of microorganisms and the formation of biofi lms. Various hydrophilic polymers are used as protein-resistant coatings. Among them, poly(ethylene glycol) (PEG) has been the most frequently investigated polymer to impart biopassivity to surfaces . Nevertheless, several limitations of PEG-based technology have been reported. One of them is loss of biopassive function in case of long-term application. It is believed that PEG degrades by auto-oxidation due to the repetitive oxygen atom in its structure (Pidhatika et al., 2012, Chen et al., 2014. An alternative polymer with a peptide-like structure is poly(2-oxazoline) (Hoogenboom, 2009). In particular, we were interested in poly(2-methyl-2-oxazoline) (PMOXA). Studies on poly(oxazoline) (POx) grafted to liposomes And micelles (Lee et al., 2003) did not report any adverse effects of the polymer in animal models thus suggesting that POx behave similar to PEG for this type of conjugates. Moreover, this polymer, with brush conformation on surfaces, was recently reported as an excellent non-fouling polymeric coating against serum-protein adsorption and bacteria adhesion (Pidhatika et al., 2010). Series of brush-like copolymers consisting of a polycationic poly(L-lysine) (PLL) backbone and PMOXA side chains (PLL-g-PMOXA) have been immobilized on Nb 2 O 5 -surface And proven to have very similar protein-repellent properties when compared to the extensively studied (PEG)based system (PLL-g-PEG) (Pasche et al., 2003). In terms of stability, PMOXA that mimics polypeptide structure was found to be significantly more stable compared to PEG with polyether structure (Pidhatika et al., 2012). However, biopassive coating may deteriorate during longer-term contact with bacteria or body fl uid in general due to mechanical defects of the layer and secretion of oxidative substances from bacteria, requiring other strategies to explore.
There have been attempts to prevent device-related infections by immobilizing bioactive antimicrobial compounds on surfaces. Quaternary ammonium compounds (QAC) are the most frequently studied antimicrobial compounds immobilized on various surfaces. Bouloussa et al. (2008) immobilized QAC on hydroxylated surfaces like glass, silicon wafers and cellulose And demonstrated that QAC could kill 99.7%-100% of Gram-negative and Grampositive bacteria. Brizzolara and Stamper (2007) covalently immobilized QAC on fl at glass substrates and 1-μm-diameter silica microparticles respectively, and investigated the antibacterial activity of these surfaces against Staphylococcus aureus (S. aureus). Klibanov and co-workers (Klibanov, 2007, Nebojscarona et al., 2005, Tiller et al., 2001 showed high activity of QASs on fl at surfaces. The antibacterial activity of QAC is induced by the mechanical contact between the ammonium cation And the anionic bacterial membrane, resulting in disruption of the membrane system (Gottenbos et al., 2002).
Here we present the development of novel dual-functional (biopassive And bioactive) biocompatible surface coatings as model systems for testing their functionality towards the prevention of biomaterial-related infections. Polyelectrolyte co-polymer brush composed of poly (2-methyl-2-oxazoline) chains grafted to poly(L-lysine) backbone (PLL-g-PMOXA) serves as a biopassive protective layer on the surface. When hydrated, it prevents the non-specific adsorption of biologically active species. Further, introduction of bioactive quaternary ammonium compound (QAC) into the biopassive copolymer brush provides dual functionality of the coating with the aim of increasing its antimicrobial effi ciency and reducing the extent of dead bacteria fouling. The coupling of the QAC to the PLL-g-PMOXA copolymer matrix is performed by copper(I)-catalized cycloaddition reaction (click reaction). This reaction proceeds in a variety of solvents, tolerates a wide range of pH values, and performs well over a broad temperature range. It is quantitative, stereospecifi c and highly tolerant towards different functional groups. Click reaction is defi ned as the Cu(I)-catalyzed cycloaddition between terminal acetylenes And azides, resulting in 1,4-disubstitution of 1,2,3-triazole (Bock et al., 2006).

Materials
Poly(L-lysine) (PLL) was purchased from Sigma-Aldrich. Initiator prop-2-ynyl-4methylbenzenesulfonate (propargyl tosylate) was purchased from Fluka and used as received. Monomer 2-methyl-2-oxazoline (MOXA) was purchased from Acros Organics and distilled under reduced pressure right before used with KOH was used as a drying agent. Terminating reagent ethyl piperidine-4-carboxylate (4-piperidine ethyl ester, PipEtEst) was purchased from Acros Organics and purifi ed by vacuum distillation. The obtained distillate was stored under argon at -4 o C. HPLC grade Acetoniltrile (ACN) was purchased from Fluka Chemicals And freshly dried by refl uxing over CaH 2 under a dry argon atmosphere. Azido quaternary ammonium compound N 3 -C 3 H 6 -N + -C 12 H 25 was provided from custom synthesis. Catalysts for carbodiimide chemistry, N-hydroxysulfosuccinimide s o d i u m s a l t ( S u l f o -N H S ) A n d A n d N -( 3 -d i m e t h y l a m i n o p r o p y l ) -N 'ethylcarbodiimide hydrochloride (EDC) were purchased from Pierce and Sigma Aldrich, respectively. Dialysis tubing was purchased from Spectrum Laboratories, Inc.
Ultrapure water was used for dialysis, rinsing step, And making buffer. Ultrapure water was paoduced in a Milli-Q system Gradient A 10 from Millipore (Zug, SwitzerlAnd). The system is equipped with Elix 3, which removes 95 -99% of inorganic ions, 99% of dissolved organic compounds, bacteria, And particulates.
4 -( 2 -h y d r o x y e t h y l ) -1piperazineethanesulfonic acid (HEPES) And other chemicals used for buffer preparation were purchased from Fluka (Buchs, SwitzerlAnd). During the entire study, HEPES solution with additional 150 mM NaCl was used. The buffer solution was adjusted to pH = 7.4 using 6 M NaOH.

Synthesis of PLL-g-PMOXA-QAC
The synthesis route of PLL-g-PMOXA-QAC is shown in figure 1. Carboxyterminated PMOXA was fi rst prepared by living cationic ring opening polymerization (Konradi et al., 2008, Von Erlach et al., 2011. The 2-methyl-2-oxazoline (8.5 ml, 100 mmol; freshly distilled from KOH through a vigreux column) was dissolved in acetonitrile (30 ml; freshly distilled from CaH 2 ) under argon. Prop-2-ynyl-4-methylbenzenesulfonate (2 mmol; used as received) was added at 0°C under argon. The mixture was then heated to 70 o C for 20 h. Subsequently, the reaction mixture was cooled to room temperature And the reaction was quenched by addition of ethyl piperidine-4-carboxylate (6 mmol; distilled through a vigreux column) for 12 h. The solvent was removed under reduced pressure to yield clear colorless oil. After addition of water (MilliQ, 18.3 MΩ·cm), the polymers were purifi ed by dialysis in a 1000 MWCO tubing for 2 × 24 h. C 3 H 3 -PMOXA-COOC 2 H 5 was collected as a white solid (4 g, ~50%) after lyophilization.
In the second step, the carboxy ethyl ester (-COOC 2 H 5 ) end group was hydrolyzed by reaction with sodium hydroxide solution (NaOH (aq) ) at pH = 14 And room temperature for 24 h, resulting in C 3 H 3 -PMOXA-COOH. The polymers were purifi ed and lyophilized using the same procedure as described above.
The C 3 H 3 -PMOXA-COOH polymers were then grafted onto PLL backbone (poly(L-lysine) hydrobromide, MW 15000 -30000 g/mol) in aqueous buffer solution of pH = 7.4 (10 mmol/l HEPES) using water soluble carbodiimide chemistry. PLL, C 3 H 3 -PMOXA-COOH, Sulfo-NHS And EDC were individually dissolved in HEPES buffer. PLL solution was then mixed with C 3 H 3 -PMOXA-COOH solution, followed by Sulfo-NHS, And fi nally EDC solution. Sulfo-NHS was   Pidhatika & Rakhmatullina against water followed by lyophilization. The molecular weight cut off was chosen such that both, low molecular weight reagents And unreacted C 3 H 3 -PMOXA-COOH were effectively removed. Alkynyl PLL-g-PMOXA was obtained as white powder with typical yields on the order of 80% (200 mg).
In the final step, the alkynyl PLLg-PMOXA copolymer (20 mg, 1 eq) was reacted with azido quaternary ammonium compound N 3 -C 3 H 6 -N + -C 12 H 25 (0.374 mg, 10 eq) using click chemistry method in a 1:1 H 2 O/t-BuOH solvent, with Cu(II) And Naascorbate as the catalysts. The click reaction was performed at room temperature for 24 hours, followed by dialysis, lyophilisation, And characterization.
The synthesized polymers were characterized using matrix-assisted laser desorption/ionization-time of fl ight (MALDI-ToF) spectrometry And nuclear magnetic resonance (NMR) spectroscopy. MALDI-ToF measurements were performed on a Bruker Daltonics Ultrafl ex II instrument, while NMR spectroscopy was performed on a 500 MHz Bruker Ultrashield Avance instrument.

Surface modifi cation
The PLL-g-PMOXA-QAC copolymers were then adsorbed onto carboxylated tissue cultured polystyrene (TCPS) surfaces (24 well plates). These surfaces expose a net negative charge, while the lysine units from the PLL backbone that are not bound to PMOXA expose a net positive charge in physiological pH. This condition results in electrostatic attachment between the TCPS surfaces And the PLL-g-PMOXA-QAC copolymers. Polymer fi lms were prepared by dip and rinse procedure. Briefly, the polymers were dissolved at 0.1 mg/ml concentration in a fi ltered (0.22 μm) HEPES buffer solution containing 10 mM HEPES supplemented with 150 mM NaCl And adjusted to pH 7.4 (HEPES2). Then, 500 μl of copolymer solutions were placed onto the freshly opened carboxylated tissue cultured polystyrene (TCPS) surfaces, completely covering their surfaces. Polymer adsorption was allowed to proceed for minimum 2 h, followed by extensive washing with HEPES2 solution, ultrapure water, and finally by blow-drying under a stream of nitrogen.

Bacteria culture
Bacteria E. coli K12 AAEC191A that express green fluorescence protein (eGFP) were grown overnight in LB media, supplemented by antibiotics, at 37°C And 100 rpm of shaking. The eGFP expression was then induced by adding IPTG with 1 mM fi nal concentration, followed by incubation at 30°C And 100 rpm of shaking, for minimal 3-4 h. The bacteria were washed with M9 minimal media supplemented with antibiotics And IPTG, and the bacteria concentration was adjusted to 10 9 bacteria/ml.

Bacteria assay And live-And-dead staining
The PLL-g-PMOXA-QAC-coated TCPS surfaces were exposed to freshly prepared bacteria solution (10 9 bacteria/ ml) at room temperature. The volume of bacteria solution was 500 μl per TCPS well. Following 1 h of exposure, the surfaces were washed thoroughly with M9 minimal media supplemented with antibiotics And IPTG, keeping the surfaces immersed in the solution at all times. Propidium iodide stain was then added to the solution after washing with 30 μM fi nal concentration. Live and dead surface-adherent bacteria were imaged using an inverted Nikon TE2000-E microscope equipped with a Hamamatsu Photonics EM CCD-C9100 camera using Metamorph software (Molecular Devices, Inc.).

Results And discussions
M A L D I -T o F s p e c t r o m e t r y w a s performed to determine the molecular weight of PMOXA. Figure 2 shows the MALDI-ToF spectrum of (C 3 H 3 )-PMOXA-(COOC 2 H 5 ) (Mn = 4400 g/mol, DP = 50, PDI = 1.01). All detected peaks presented a peak-to-peak average distance equal to 85 g/mol, which confi rmed the repeating Pidhatika & Rakhmatullina unit of 2-methyl-2-oxazoline (MOXA). The main peaks show the molecular weight of the corresponding chemical species coupled with a sodium ion.
In addition to the main peaks, second minor peaks with lower masses relative to the main peaks at 38-mass differences were detected. For MALDI-ToF spectra of poly(2-methyl-2-oxazoline) and poly(2-ethyl-2-oxazoline), this phenomenon has been observed earlier and could be attributed to an initiation of a small fraction of monomers with a proton (H + ) instead of the initiator (Gaertner et al., 2007, Jordan et al., 2001. Thus, a minor part of the polymer presents as (H)-PMOXA-(COOC 2 H 5 ) instead of (C 3 H 3 )-PMOXA-(COOC 2 H 5 ). However, the intensity of this minor distribution is low and the molar mass of the resulting polymer is not signifi cantly affected.
Proton ( 1 H) NMR spectroscopy was performed to confi rm the chemical structure of the synthesized polymers And graft copolymers. Figures 3a shows the NMR spectra of (C 3 H 3 )-PMOXA-(COOC 2 H 5 ), before and after hydrolysis of the ester group.
The (C 3 H 3 )-PMOXA-(COOH) was then coupled to a PLL backbone presenting -NH 2 along its chain, by using carbodiimide chemistry in aqueous solution. NMR spectra of alkynyl PLL-g-PMOXA is presented in fi gure 4. Upon grafting (amide formation), changes of peak intensities were observed in the chemical shifts at 4.2, 3.05, 2.9, and between 1.2 -1.8 ppm. Grafting density α was calculated from the ratio between peak intensity of -CH 2 -groups adjacent to amide And amine, respectively. Thus, α = I -CH2-N(C=O)-   , which results in α value of 0.33. We have previously found that this value is the optimum grafting density value for PLL-g-PMOXA brush as biopassive surface coating (Pidhatika et al., 2010, Pidhatika et al., 2012. Figure 5 shows the NMR spectrum of PLL-g-PMOXA-QAC and the assignment of peaks. The successful conjugation between PLL-g-PMOXA and QAC could be seen by the appearance of QAC peaks in addition to the PLL-g-PMOXA peaks, although some overlap were observed.

Pidhatika & Rakhmatullina
T h e i d e a l c o n f o r m a t i o n o f t h e copolymers on the surface is illustrated in fi gure 6. The anchor (PLL backbone) serves positive charges from the lysine groups that are not connected to the PMOXA side chains while the surface (TCPS) serves negative charges. The PLL-g-PMOXA-QAC copolymer is thus immobilized on the TCPS surface by electrostatic interactions. The bioactive part (QAC) serves as antimicrobial compound that kills (lyses) approaching bacteria cells to the surface. The biopassive part (PMOXA) serves as anti-adhesive layer that repels microorganisms And biomolecules including the residues from dead bacteria And the lysate.
The dual-functional antimicrobial surface was then exposed to bacteria solution in order to test its antimicrobial properties. PLL-coated And bare TCPS surfaces served as controls. Figure 7 shows the microscopy images after live-And-dead staining assays on the surfaces. Green cells indicate the presence of living bacteria, whereas red cells indicate dead bacteria on the surface.
It is observed in figure 7 that PLLcoated surface was fully covered with living bacteria. At physiological pH, PLL-coated surface presents positive charges originating from the positively charged free lysine units, causing electrostatic interactions with negatively charged bacteria cells (Pidhatika et al., 2010). In contrary to PLLcoated surface, bare carboxylated TCPS surface presents negative charges, thus a decrease in number of adherent bacteria was observed due to repulsive interaction between similarly (negative) surface charges.
However, several living bacteria managed to adhere on the bare surface due to the absence of steric stabilization and entropic penalty to inhibit strong, irreversible interactions. Interestingly, PLL-g-PMOXA-QAC-coated surface shows very few dead bacteria. This phenomenon indicates the dual-functionality of the surface, i.e. the bioactive compound  presumably kills (interferes) the bacteria approaching the surface, and the biopassive layer repels both living bacteria and protein residue from dead bacteria.

Conclusions
We have successfully synthesized graft copolymer PLL-g-PMOXA (α = 0.33) with acetylene functionality as a biopassive platform in the development of dualfunctional antimicrobial surfaces. The acetylene functional groups served as the binding sites for the conjugation between the PLL-g-PMOXA And bioactive QAC using click reaction. The NMR spectrum of the resulting PLL-g-PMOXA-QAC indicated successful conjugation between PLL-g-PMOXA and QAC. The successful conjugation was further confi rmed by the results of preliminary dualfunctionality test by means of live-and-dead bacteria assay.