Preventing Pseudomonas aeruginosa Biofilms on Indwelling Catheters by Surface-Bound Enzymes.

In our previous study, we covalently immobilized PslG h to flat glass, PDMS and PS squares, to demonstrate that a continuous enzymatic degradation of Psl disrupted P. aeruginosa attachment and biofilm formation 51 . Extending this approach to medically relevant devices, we first tested conditions for uniform immobilization of PslG h to the lumen surface of catheters of reasonable length (>10–20cm), where physical access is difficult. The materials of commercial catheters, PE, PDMS and PU, are useful for their chemical inertness (in addition to low friction coefficient, strength, and sterilizability), but also make them difficult to chemically functionalize. For PslG h immobilization, we used a 4-step process: surface hydroxylation by plasma oxidation; amine-functionalization by APTMS; GDA cross-linking and PslG h covalent binding ( Figure 1A ) 51 . Surface hydroxylation is the key step in PslG h immobilization to generate active sites (e.g. OH and COOH) for APTMS 57 . The plasma treatment time to hydroxylate PE surface was tested first for simple PE sheets (1cm 2 ), and their surface wettability was evaluated by water contact angle. Treating PE sheets with 5 min plasma exposure significantly reduced the contact angle from 92° ±2 (hydrophobic) to 21° ±1 (hydrophilic) ( Figure S2A ), indicating an increase in surface polar groups. ATR-FTIR confirmed a weak O-H broad band at ~3400 cm −1 ( Figure S2B ). The low intensity of the O-H signal is due the thin surface layer and the limited sensitivity of ATR-FTIR 57 . For PDMS and PU sheets, atmospheric plasma treatment for 5 min also caused a reduction in contact angle from 108° ±4 and 69° ±2 to 21° ±3 and 20° ±1, respectively. Tubing segments (0.5 cm, ID) of PE, PDMS and PU were then treated with plasma for 5 min, silanized with APTMS and the PslG h cross-linked with GDA. The success of chemical immobilization along tubing length was confirmed by ATR-FTIR after each step ( Figure 2A – C ). ATR-FTIR spectrum of the lumen surface of untreated (control) PE tubing ( Figure 2A ) showed three ethylene bands at 3000–2800 cm −1 (CH 2 asymmetric and symmetric stretching), 1500–1400 cm 1 (1473 and 1463 cm −1 bending and wagging deformation at 1366) and 730–720 cm −1 (rocking). After plasma treatment, we observed a small O-H broad band centered at ~3400 cm −1 of the hydroxyl (OH) groups 56 . Addition of APTMS shows FTIR spectra with peaks belonging to C-H (~2900 cm −1 ) and N-H (~3300 cm −1 ) stretching vibrations, and peaks corresponding to hydroxyl groups (~3400 cm −1 ) and C=O (~1635 cm −1 ) stretching vibrations after the addition of GDA. After incubating with PslG h , the spectrum showed a signal at 1635 cm −1 associated with the carbonyl (C=O) vibrations of an amide I band, and a broad band of O-H and N-H stretching vibrations centered at 3343 cm −1 . These signals were similar to those obtained from the IR analysis of pure PslG h in buffer solution 51 . For PDMS and PU tubing, similar modifications in the surface chemistry were detected by ATR-FTIR ( Figure 2B and C ) confirming the attachment of PslG h to these two surfaces. Collectively, these surface characterizations confirm successful covalent immobilization of PslG h to the lumen surface of PE, PDMS and PU tubing. After PslG h immobilization on wide tubing (0.5 cm, ID), we next tested short segments (2.5 cm) of medical grade PE-100 microcatheter (0.8 mm, ID). As shown in Figure 2D , we could measure changes in lumen hydrophobicity by the relative capillary rise of water due to capillary action 58 . The capillary height for the control PE-100 catheters was ~ 1 cm below the water level ( Figure 2D ) due to its hydrophobic surface. Plasma treatment increased the water level by ~ 2 cm, showing a hydrophilic character. Adding APTMS, the water level decreased below the surface level due to the increased number of hydrophobic C-H and N-H groups on the surface ( Figure 2D ), while GDA increased the capillary rise again due to the hydrophilic hydroxyl (O-H) and carbonyl (C=O) groups. After PslG h immobilization, the lumen surface maintained its hydrophilic character. The average length of a CVC catheter insertion is 11–14 cm 9 , so for the in vitro and in vivo -testing we prepared 14 cm lengths of PE-100 medical grade tubing with PslG h . As the lumen of a longer catheter is not directly exposed inside the plasma treatment chamber, the efficiency of hydroxylation may be affected. To test the uniformity along the tubing length, the catheters were cut into 4 equal segments after each immobilization step, and their hydrophobicity and surface chemistry measured by capillary rise and ATR-FTIR, respectively. We found that plasma treatments of < 10 min resulted in variations in the level of water reached within each segment (S1-S4) of the lumen, but exposures of 10 min or more resulted in uniform hydroxylation with even capillary rise for all 4 segments ( Figure 2E ). Importantly, identical water level (indicative of hydrophilicity) ( Figure S3A ) and ATR-FTIR spectra (indicative of surface chemistry) ( Figure S3B ) were also observed for all 4 sections after PslG h immobilization, suggesting that PslG h was also evenly immobilized along the entire length of these PE-100 catheters. We first tested the resistance of covalently-bound PslG h on PE, PDMS, PU tubing to P. aeruginosa PAO1 biofilm formation under static culture conditions. Fluorescence microscopy using SYTOX green was used to visualize attached bacterial cells and biofilm formation after 24 h growth. While dense biofilms were observed on the three untreated controls, those with bound PslG h showed significantly lower cell attachment and colony formation ( Figure 3A ). Image analysis (cell count/cm 2 ) revealed that the bound PslG h significantly reduced the number of attached cells by at least 2-log ( Figure 3B ). These results confirm the uniformity of bound PslG h along the lumen length of PU and PDMS and that the surfaces can effectively inhibit the adhesion, colonization and subsequent biofilm formation of P. aeruginosa PAO1. Previously, we found that the enzymatic activity of covalently-bound PslG h on glass surfaces was dependent on the enzyme surface density 51 . To maximize the immobilized PslG h surface density on the lumen surface of PE-100 microcatheters, we varied the PslG h concentration (0, 20, 40, 80 and 160 μg/ml) in the immobilization buffer. The surface density of bound PslG h was determined from the difference between its concentration before and after immobilization using the Bradford protein assay 51 . The amount of surface-bound PslG h increased exponentially with increasing PslG h concentration in the buffer up to a PslG h concentration of 80 μg/ml, at which point the maximum immobilization capacity is likely reached ( Figure 4A ). The maximum PslG h surface density was estimated to be about 1.86 μg/cm 2 . Fluorescence microscopy images showed a thick biofilm developed on the lumen surface of short (2.5 cm) untreated PE-100 microcatheter after exposure to P. aeruginosa PAO1 under static culture condition for 24 h ( Figure 4A ). However, a moderate reduction in the biofilm growth was observed even at a low density (1.18 μg/cm 2 ) of surface-bound PslG h , and complete inhibition of bacterial attachment and biofilm formation was observed at a PslG h surface density of 1.78 μg/cm 2 and above ( Figure 4A ). To ensure that the anti-biofilm activity was due to the bound PslG h specifically, we tested the effects of surface functionalization with APTMS and GDA on P. aeruginosa biofilm formation. The number of bacterial cells attached to APTMS (1.7 ×10 6 CFU/cm 2 ) and GDA (2.7 ×10 6 CFU/cm 2 )-treated lumen surfaces after 24 h were not significantly different from that of the untreated microcatheter lumen (2.3 ×10 6 CFU/cm 2 ). However, a ~3-log reduction in attached cell numbers was observed with PslG h -immobilized vs untreated microcatheters (1.7 ×10 3 vs 2.3 ×10 6 respectively, Figure 4B ). In addition, SYTOX green (SG) and crystal violet (CV) staining demonstrated the formation of thick biofilm on the lumen surface of untreated microcatheters and those treated with APTMS or GDA, while biofilms were absent from the lumen of PslG h -treated PE-100 microcatheters ( Figure 4B ). Combined, these results demonstrate that the anti-biofilm activity of the PE-100 microcatheters against P. aeruginosa is due to the immobilized PslG h rather than other effects of the immobilization process on the microcatheter lumen surface. To test the activity of the modified PE-100 microcatheters against clinical strains, we used P. aeruginosa strain ATCC 27853, a clinical strain with great ability to produce a more robust biofilm compared to the lab PAO1 strain 59 . We first tested the ability of PslG h in solution to inhibit its biofilm formation. Our results demonstrate that the PslG h in solution is effective at inhibiting P. aeruginosa ATCC 27853 biofilms with an EC 50 of ~ 1 nM ( Figure S4A ). Fluorescence microscopy images of the lumen surface of untreated PE-100 microcatheter showed a thick biofilm after exposure to P. aeruginosa ATCC 27853 under static culture condition for 24 h ( Figure S4B ). However, a significant reduction in the number of attached cells (~3-log) was observed compared to the control ( Figure S4C ).To ensure the activity of immobilized PslG h along the entire microcatheter length (10 cm), we tested the anti-biofilm activity against P. aeruginosa PAO1 biofilm under 24 h static culture. After incubation, CV staining of biofilm mass along the 10 cm microcatheter showed almost no biofilm formation on the PslG h -bound lumen surface, whereas a heavily CV-stained biofilm appeared within the untreated PE-100 control catheter after 24 h ( Figure 5A ). To test the uniformity of the anti-biofilm activity along the entire 10 cm catheter lumen, catheters were cut into ten 1-cm long segments after incubation in the same conditions as above. Then, 5 alternating segments ( Figure 5B ) were tested for resistance to P. aeruginosa PAO1 biofilm formation using fluorescence microscopy (S2, S4, S6, S8, S10) and CFU (S1, S3, S5, S7, S9). Microscopic images showed thick biofilms within all 5 segments of the untreated catheter lumen ( Figure 5C ). In comparison, the covalently-bound PslG h significantly eliminated biofilm formation of P. aeruginosa for all PE-100 segments ( Figure 5C ). Quantitative culture confirmed that the PslG h -treatment of catheters significantly reduced attached cell density from 6.5–6.8 ×10 6 CFU/cm 2 (untreated control) to 1–8.3 ×10 3 CFU/cm 2 (PslG h -bound) after 24 h, representing a ~3-log reduction in CFU ( Figure 5D ). These results confirm we succeeded in generating a uniform, high density PslG h coating of the lumen surface, with strong anti-biofilm activity against P. aeruginosa along the length of these medical grade PE-100 catheters. We next tested the PslG h -modified catheters in an in vitro flow culture model ( Figure S1 ). P. aeruginosa cells were allowed to adhere under static culture conditions for 1 day after which a constant flow (0.6 mL/min) of fresh LBNS media was passed through the lumen of both untreated control and PslG h -bound PE-100 catheters in parallel. Fluorescence microscopy ( Figure 6A ) demonstrated that the PslG h -bound catheters were resistant to biofilm formation on the lumen surface for the 14 d test period under dynamic culture flow conditions, as compared to the control PE-100 catheters. The rate of cell attachment and biofilm formation was also measured under these flow culture conditions. Quantitative culture confirmed that PslG h -bound catheters exhibited an ~ 3 log reduction in attached cell numbers during the first 11 d of flow, as compared to untreated catheters. Bacterial counts for the untreated catheters decreased between days 11 and 14, which resulted in a reduction of only 2 log between the treated and control catheters at day 14 ( Figure 6B ). Taken together, these results indicated that PslG h immobilized on the catheter lumen surface inhibits biofilm growth for at least 14 d under dynamic flow culture conditions. As a prelude to an in vivo study, we investigated the effect of blood proteins, storage stability and catheter sterility on the antibiofilm activity of the PslG h -bound catheters. Immediately after catheterization, the surface of indwelling catheters becomes covered with plasma proteins that could potentially interfere with the function of the immobilized enzyme. Therefore, the activity of the lumen-bound PslG h was tested after being treated with serum for 24 h. In general, there were significant increases in the number of cells that adhered on both untreated and treated catheters after serum treatment ( Figure S4 ). Nevertheless, CV staining ( Figure S4A ) and SYTOX green ( Figure S4B ) both confirmed a reduction in bacterial biofilm biomass on the lumen surface of PslG h -bound catheters relative to the untreated control catheter, despite treatment with serum. A few cell clusters were observed by fluorescence microscopy of the PslG h -bound catheter lumen ( Figure S4B ). Quantitative culture showed that, relative to the control catheters, immobilized PslG h reduced attached cell density within the catheter lumen before and after serum treatment by 3-log and ~2.5-log, respectively, ( Figure S4C ). Our results show that while serum proteins increase the adherence of cells to both the control and enzyme-treated surfaces, the PslG h -bound PE catheters remain more resistant to cell attachment even in the presence of serum. Immobilization procedure ( Figure 1 ) was carried out under aseptic conditions. Prior to immobilzations, caatheters were either pre-sterilized using70% ethanol or ethylene oxide. The immobilzation steps involve the use of atmospheric plasma, 80% ethanol, 2% GDA and filter sterilized PslG h in PBS buffer which are sufficient to maintain the sterility of catheters during immobilization. To confirm, we tested the sterility of PslG h -bound catheters with a standard plate counting method. Results confirmed the absence of contamination. In addition, we tested the storage stability of PslG h -bound catheters after storage for 30 d at 4 °C under both dry and wet conditions. PslG h -bound catheters were found to remain resistant to biofilm formation as shown from fluorescence microscopy ( Figure S6A ) and their corresponding cell counts, which exhibited a 2.5–3 log reduction ( Figure S6B ). Finally, we tested the anti-biofilm activity of these PslG h -bound PE-100 catheters in the complex environment found in vivo , using a central venous catheter infection (biofilm) rat model 65 . In this model, the clinical strain P. aeruginosa ATCC 27853 was chosen as it was isolated from a hospital blood specimen and has been shown to produce a more robust biofilm compared to the lab PAO1 strain 59 . Twenty-four hours after intravascular placement, SEM imaging of untreated catheter segments ( Figure 7A ) showed a dense biofilm network that consists of a relatively large number of cells enclosed in an extracellular matrix. In comparison, segments from the PslG h -bound catheter contained scattered and weakly attached cells and a lack of biofilm matrix. Quantitative culture of the PslG h -bound catheter showed a significant inhibition of biofilm formation 24 h after infection, as compared to the untreated catheters (~2 log CFU reduction) ( Figure 7B ).

Aminopropyl trimethoxysilane (APTMS), glutaraldehyde (GDA), ethanol (HPLC grade), Fetal bovine serum (FBS) and all other reagents were of high analytical grade and were purchased from Sigma Aldrich (Mississauga, ON). Commercial polyethylene (PE), polyurethane (PUR) and polydimethylsiloxane silicone (PDMS) tubing with identical dimensions (0.3 cm ID and 0.6 cm OD) from McMaster-Carr (Ohio, USA), and medical grade polyethylene (PE-100) catheters (0.8 mm, ID) (Intramedic, Becton Dickinson and Company, NJ) were used as biomaterial surfaces. Strains used in this study are P. aeruginosa PAO1 (wild-type strain; serotype O5) and P. aeruginosa ATCC 27853 (clinical strain). Lysogeny broth (LB) and agar were used for standard culture conditions. To prepare a pre-culture of tested strains, a single colony of freshly grown bacterial cells grown on LB agar overnight at 37 °C was inoculated in 5 mL LB and incubated at 37 °C overnight with agitation. Bacterial biofilms were grown in LB without salt (LBNS). Serum treatment was performed by coating the catheter segment with 100 μL of FBS then incubating without shaking at room temperature for 30 min. The hydrolase domain of PslG h was cloned into an expression plasmid and purified from E. coli as previously described 50 , 52 . Briefly, the enzyme was recombinantly expressed in E. coli cells and grown in autoinduction media overnight at 37 °C with 50 μg/mL kanamycin. The protein was purified by Ni-NTA affinity chromatography using an N-terminal His 6 -tag present on the enzyme followed by size exclusion chromatography (HiLoad 16/600 Superdex 200 pg, GE Healthcare). The protein was eluted in 20 mM Tris (pH 7.5), 150 mM NaCl, and 2% (vol/vol) glycerol and stored at −80 °C until required. The His 6 -tag was left intact for all experiments. Enzyme immobilization by covalent binding for the activated surfaces was performed as previously described 51 . We took advantage of surface hydroxyl or carboxylic groups as reactive sites for the APTMS, creating primary amines to be covalently linked to those of the PslG h by GDA ( Figure 1A ). All polymer sheets and tubing (PE, PU and PDMS) and medical grade PE catheters (PE-100) were sonicated in absolute ethanol for 15 min then oven dried. Dried samples were exposed to atmospheric plasma (Harrick Plasma PDC001) at high power for up to 10 min. To ensure an even hydroxylation of the lumen surface, sections were flipped end to end after 5 min then exposed to plasma for additional 5 min. Samples were then functionalized with amine (NH 2 ) groups by immersion in APTMS (Sigma-Aldrich, 97%) solution (50 mg/mL in 80% (v/v) ethanol) for 2 h with shaking at 25 °C according to a published protocol 53 . Tubing segments were checked for any air bubbles that may prevent the APTMS contact with the lumen. After silanization, the surfaces were washed three times with 80% (v/v) ethanol to remove unreacted APTMS. The amino-functionalized surfaces were then immersed in 2% (v/v) GDA in 1×PBS buffer solution (pH 7.2) for 2 h under gentle stirring at 25 °C. The surfaces were rinsed 3 times with sterile 1× PBS to remove unreacted GDA. Functionalized (NH 2 -GDA) surfaces were immersed in PslG h solution (80 μg/mL in 1x PBS buffer) and incubated overnight at 4 °C then washed several times with 1× PBS buffer to remove the unbound enzyme. The concentration of immobilized enzyme was determined from the difference between the concentration of PslG h in solution prior to immobilization minus the amount of protein that remained in the solution after the immobilization and washings steps. The enzyme concentration before and after immobilization and in washing buffers was measured using the Bradford protein assay (B6916, Sigma) with bovine serum albumin as a standard. Surface wettability measurements and attenuated total reflection Fourier transform infrared spectroscopy (ATR-FTIR, PerkinElmer Spectrum 100) were used to monitor each step of the process including the surface activation (hydroxylation), silanization (APTMS), cross linking (GDA), and enzyme covalent binding (correlated with the spectra of native components). After a background scan, spectra were collected from 4000 to 625 cm −1 using 64 scans at 4 cm −1 resolution. The reflection intensities were corrected to compensate for their frequency (wavenumber) dependence (i.e., neglect the effect of frequency). Surface wettability of ‘wide’ tubing segments was assessed by measuring the static water contact angle using the sessile drop method. Tubing segments were cut longitudinally into two halves and a 1 μL drop of dH 2 O was placed on the lumen surface and imaged. The angle between the horizontal plane and the tangent to the drop at the point of contact with the surface was analyzed by ImageJ 54 using the contact angle plugin 55 . For thin PE-100 catheters, the lumen surface wettability was evaluated by lowering the catheters into identical vials half filled with distilled water and measuring the water level change inside the lumen. To evaluate the surface chemistry of the lumen, tubing was cut longitudinally into small, low curvature segments and the interior surface (lumen) was pressed firmly against the face of the ATR crystal to improve surface contact and obtain high quality absorbance/transmission signals. PAO1 cells were grown in LB medium at 37 °C overnight with shaking at 200 rpm. Cell cultures were normalized to an OD 600 nm of 0.5 and diluted in LBNS (1:100). 5 mL of the diluted culture was added to sterile 6-well polystyrene microplates (Thermo Scientific cat. no. 243656) containing the tested catheters segments (2.5 cm). The culture plates were incubated for 24 h at 25 °C. After incubation, catheters were removed, drained, and rinsed 3x with 10 ml of sterile 1× PBS buffer (pH 7.2) to remove non-adherent cells and media. Under strict sterile conditions, the catheters were cut into 1 cm segments and transferred to separate sterile vials for Sytox Green, crystal violet (CV) staining and colony forming units (CFU) counts, respectively. P. aeruginosa PA01 biofilms were grown in a lab-made ‘two catheters’ device, in which the sterile untreated (control) and PslG h bound PE-100 (10 cm) were installed together with one inlet and one outlet device ( Figure S1 ). P. aeruginosa PAO1 was grown in LB Broth overnight, with shaking at 37 °C. The culture was then diluted 1:100 into 150 mL of fresh LBNS Broth in a 250 mL Erlenmeyer flask and stirred. The device inlet was connected to the inoculated media, and the outlet was connected to a tube mounted into a peristaltic pump (Cole Parmer) head and into a waste media bottle. The experiment began once the inoculated medium started flowing from the end of PE-100 catheters, and the flow was stopped to allow for cell attachment and biofilm growth under static conditions. After 24 h, under aseptic conditions the Erlenmeyer flask was disconnected and 20 L of diluted (1/10) and sterilized LBNS media in a container was connected to the device and media flow (0.4 mL/min) was started. LBNS media was pumped through the reactor from the top, and the waste was dispensed from the side to ensure removal of all free, unattached cells and allow the attached cells to form biofilms under flow conditions for 14 d. A filtered port was used to provide air to the bacteria. A 1 cm section of each tubing was taken daily under sterile conditions and used in a crystal violet (CV) assay, and for SYTOX Green and CFU measurements. For PE catheters, multiple 10 cm sections of the control and PslG h -bound were connected between two multichannel devices. In this case, after long-term continuous flow (~14 d), a whole catheter was disassembled and cut into sections under sterile conditions for biofilm analysis (CV, fluorescence microcopy and CFU). To examine biofilm inhibition using a microtiter plate assay, P. aeruginosa strain ATCC27853 was grown at 37 °C overnight with shaking at 200 rpm. The cultures were normalized to an OD 600 of 0.5 and then diluted 1:100 in LBNS. Diluted culture was added to sterile 96-well polystyrene microtiter plates (Thermo Scientific cat. no. 243656), and varying concentrations of PslG h (0.1 to 10 μM) were added. The cultures were incubated statically for 24 h at 37 °C to allow for biofilm formation. To eliminate edge effects, ~200 μL of sterile water was placed in all outside wells, and the plate was sealed with parafilm. After incubation, nonadherent cells and media were removed by thoroughly washing the plate with deionized water. The wells were stained with 150 μL of 0.1% (w/v) crystal violet for 10 min and then rinsed with water. The remaining dye was solubilized by addition of 150 μL of 95% (v/v) ethanol and left for 10 min, after which the absorbance was measured at 595 nm using a SpectraMax M2 spectrophotometer (Molecular Devices). The amount of biofilm was proportional to the absorbance from staining with crystal violet. Experiments were performed in triplicate and all statistical analysis were performed using GraphPad Prism. P. aeruginosa ATTC 27853 biofilm formation on coated catheters in vivo was evaluated using a rat venous catheter model previously developed to study Candida albicans biofilm formation 63. Pathogen-free female Sprague-Dawley rats weighing 400 g (Harlan Sprague-Dawley, Indianapolis, Ind.) were used. Animals were maintained in accordance with the American Association for Accreditation of Laboratory Care criteria, and all studies were approved by the institutional animal care committee. Briefly, 24 h after implantation, the jugular venous catheters (PE-100) were inoculated (10 7 cells/mL) and allowed to dwell for 6 h before washing excess inoculum. Rats were sacrificed after 24 h and the catheters collected for bacterial enumeration or imaging by scanning election microscopy (SEM) 56 . To quantify viable bacteria as colony forming per units (CFU), catheters segments were immersed in 10 mL of 1× PBS, sonicated at low power for 5 min and vortexed for 15 s to release viable cells. Aliquots were diluted 10 2 -to10 8 -fold, and 100 μL of each dilution was plated in duplicate on LB plates. The number of colonies was counted after incubation at 37 °C for 24 h. After catheter segments were rinsed with 1×PBS at pH 7.2, the adherent bacteria were fixed by 2.5% (v/v) glutaraldehyde (GDA) in 1×PBS solution for at least 1 h, then rinsed 3 times with 1×PBS to remove unreacted GDA. After fixation, samples were treated with Tween-20 (0.05% (v/v) in PBS buffer) for 10 min, rinsed with sterile PBS, and stained in the dark with SYTOX green (Life Technologies, CA, USA) for 30 min 37 , 56 , 62 . Samples were imaged by fluorescence microscopy (Olympus BX63, Tokyo, Japan) using a 20X dry objective. The green fluorescence of the biofilm was observed through a GFP filter (λ ex /λ em 395/470 nm). The fluorescence densities of biofilms were quantified using computerized image analysis with Olympus CellSens micro imaging software. Each data point represents the average of six images. Error bars represent standard deviation from the mean ( n = 6). Crystal Violet (CV) staining was used to visualize the biofilm biomass on catheters. Briefly, after the catheters were rinsed with 1× PBS buffer and air dried for 1 h at 25 °C, they were stained with 300 μL of 0.5% (w/v) CV for 30 min, then rinsed with dH 2 O. CV-stained catheters segments were air dried and photographed. Catheters segments were rinsed with 1× PBS buffer and placed in fixative (1% (v/v) GDA and 4% (v/v) formaldehyde) overnight. The samples were rinsed in 1× PBS buffer two times and then placed in 1% (v/v) osmium tetroxide for 30 min, followed by immersion in hexamethyldislilazane (Polysciences, Inc., Warrington, Pa.). The samples were subsequently dehydrated in a series of ethanol washes (30% (v/v) for 10 min, 50% for 10 min, 70% for 10 min, 95% for 10 min, and 100% for 10 min). Final desiccation was accomplished by critical-point drying (Tousimis, Rockville, MD). Specimens were mounted on aluminum stubs and sputter coated with gold. Samples were observed in a scanning electron microscope (Hitachi S-5700) in the high-vacuum mode at 10 kV. The images were processed for display by using Adobe Photoshop version 5.0.

Developing robust approaches to limit bacterial biofilm formation on biomaterials through non-fouling (non-biocidal) mechanisms remains relatively new for antimicrobial biomedical devices. Antimicrobial strategies that do rely on biocidal drug release 60 are generally ineffective over time as they cannot prevent surface colonization (new cells adhering to attached dead cells), as evidenced by the high rates of biofilm-related disease that persist globally 4 , 39 . Since Psl adsorption plays a key role in the mechanism of P. aeruginosa cell attachment and biofilm formation, our approach specifically targets and disrupts this microbial adhesion mechanism. This approach differs from non-specific mechanisms to non-fouling such as SLIPS (slippery liquid infused porous surfaces) 61 – 62 , and other non-wetting surfaces. For well-known, common pathogens it seems advantageous to target cell attachment, at the early stages of biofilm development. It has also been reported that, in mixed species biofilms, Psl can also contribute to the protection of other pathogens such as Escherichia coli and Staphylococcus aureus from antimicrobials 15 . Therefore, specific, selective prevention of P. aeruginosa biofilm formation should not only help enable antibiotic therapies to eradicate infections and reduce the development of acquired resistance, but also help the host immune system to eliminate infections associated with medical devices that are caused by both biofilm and non-biofilm producers. The immobilization of the PslG h glycoside hydrolase uniformly along the full length of these medical grade, small-diameter PE catheters, has been a significant technical achievement due to the inert nature of PE, and the physically confinement of access to the lumen surface (inside a 1mm tube, 14 cm long) for the plasma and solution penetration. The lumen surface of implanted catheters can become exposed to blood and body fluids, and has a high risk of biofilm infection. Previous efforts to functionalize catheter lumen surfaces, such as uniform activation with polar groups, was shown to be a challenge 63 . The narrow PE-100 catheters used herein thus represent a kind of ‘worst case’ application for medical devices, and our results suggest that a wide range of device sizes, material types and morphologies should be possible for enzyme immobilization. A key finding in our work here is stability . The results show a consistent, stable activity of these enzymes to disrupt biofilm formation (through continuous Psl polysaccharide degradation) under in vitro flow conditions for extended periods (14 d), after extensive sterilization and storage (30 d, 4°C), and in the complexity of an in vivo environment. These results suggest that PslG h remains active after surface immobilization, which may make them suitable for a wide range of medical devices. Importantly, the immobilization procedure helped to maintain sterility of the catheters as conventional methods such as ethylene oxide did not work as the sterilized PslG h surfaces lost their anti-biofilm activity ( Figure S5 ). The difference between the CFU log reduction observed in vivo ( Figure 7B ) and that observed in the in vitro assay ( Figure 6B ) likely reflects the enhanced biofilm formation capacity of strain ATCC 27853 59 as well as the effect body fluids and blood serum on promoting the biofilm growth 64 . Overall, our work demonstrates that covalent immobilization of enzymes on biomedical tubing surfaces can represent an effective and feasible design strategy for non-fouling antimicrobial material design. These represent a ‘bio-active’ interface to specifically prevent bacterial adhesion mechanisms through continuous hydrolytic activity of a biofilm matrix component.

In summary, we have generated a sterile, uniform, and high-density surface of immobilized PslG h enzyme throughout the lumen of long, small diameter, medical grade catheters, in an effort to improve the performance of indwelling catheters against microbial biofilm growth. The catheters were highly effective at preventing P. aeruginosa biofilm growth under both static and dynamic flow conditions in vitro , and even for long periods of time (11 days). Finally, preliminary in vivo results show the PslG h -bound catheters dramatic resistance to biofilm growth in the complex environment of an infection model. These results demonstrate for the first time that ‘bio-active’ enzyme-modification of surfaces can realistically be applied to a wide range of biomaterials, uniformly over large areas, and with long-term activity. There is a wide range of biomedical devices that may benefit from these effects on P. aeruginosa biofilm growth, to significantly impact the problem of hospital acquired infections, complications and costs 54 .

Biofilms formed on implanted medical devices are responsible for 50–70% of all hospital acquired infections, and can be tremendously difficult to treat 1 – 4 . Biofilms are surface-attached microbial colonies embedded within a self-produced extracellular matrix, composed of exopolysaccharide, extracellular DNA and proteins 5 . Annually, there are an estimated 14 million biofilm infections in the US, causing at least 350,000 deaths 4 . The vast majority of these infections (86–95%) are associated with microbial colonization of catheters e.g ., central-line associated bloodstream infections (CLABSI), catheter-associated urinary tract infections (CAUTI), and ventilator-associated pneumonia (VAP) 6 – 7 . The annual cost of the top five hospital-acquired infections is ~$9.8 billion in the US alone 8 . Indwelling central venous catheters (CVC) are routinely used to administer medications, blood products, and nutritional fluids to patients, for both temporary and long-term use 9 . The clean surfaces of newly-implanted catheters are highly susceptible to microbial colonization 10 – 13 . The exopolysaccharide component of the matrix impairs antibiotic penetration 14 – 15 and provides a barrier against phagocytosis by host immune cells 16 . Within biofilms, microbes are up to 1000 times more tolerant to antimicrobials than planktonic cells 1 , 10 , 17 – 18 . As a result, device-related biofilm infections are difficult to treat using systemic antibiotics 19 , and often require device extraction and replacement. Medical biofilms can involve a wide variety of organisms, but the bacterium Pseudomonas aeruginosa accounts for ~28% of device-related infections 20 – 23 . P. aeruginosa is an ubiquitous, Gram-negative, opportunistic pathogen associated with a wide array of life-threatening acute and chronic nosocomial infections, particularly in patients with compromised host-defense mechanisms 24 . P. aeruginosa infections are becoming more challenging to treat due to the emergence of multidrug resistant strains 14 – 15 , 25 – 28 . P. aeruginosa produces three distinct biofilm exopolysaccharides: Psl, Pel, and alginate 29 – 31 . The cationic Pel and neutral Psl polysaccharides are produced in varying amounts in different strains 32 , while the anionic alginate is synthesized exclusively by mucoid strains in the lungs of cystic fibrosis patients 29 – 30 . Both Pel and Psl play a significant role in the formation and maintenance of biofilm structure 14 , 33 , and recent work has shown the role that polysaccharides (such as Psl) play to enable the cell-cell and cell-surface adhesion necessary for biofilm formation 14 , 16 , 32 , 34 – 35 . Antimicrobial coatings for medical devices often incorporate silver, chlorhexidine, or triclosan for diffusive release 13 , 36 – 38 . These coatings are often ineffective over time as they do not directly prevent microbial attachment, and new cells eventually accumulate on dead cells 39 . Diffusional release is also limited by a number of factors, including: (i) slow diffusion rates to the material surface; (ii) decreased activity over time as the antimicrobial concentration decreases and bacterial biomass accumulates 40 ; and (iii) antimicrobial tolerance 41 – 42 . These limitations have led to an alternative trend to develop non-adhesive, non-fouling surfaces that target the early stages of biofilm development. These materials are designed to keep bacteria in a planktonic (swimming) state 43 – 45 and thus more susceptible to antibiotics and the host immune system 10 . Non-fouling, ‘bio-passive’ polymer brush surfaces, such as hydrophilic polyethylene glycol (PEG) are effective, but tend to fail after a period of days 46 – 47 . S lippery l iquid i nfused p orous s urfaces (SLIPS) show highly-effective, long term anti-fouling properties but clinical performance is still unclear 48 . Recently, we developed novel ‘bioactive’ biomaterial surfaces by covalently immobilizing the glycoside hydrolase, PslG h . In solution, this enzyme has been shown to disrupt P. aeruginosa biofilms in vitro 49 – 50 . We immobilized PslG h to flat squares of glass, polydimethylsiloxane (PDMS) and polystyrene (PS), and demonstrated that the surfaces prevented adsorption by hydrolyzing Psl 51 . The PslG h -bound surfaces reduced P. aeruginosa biofilms by 3–4 log units (8 d in static culture), compared to controls. However, to be suitable for clinical application, it is important to uniformly immobilize PslG h along the length of the (inner) lumen surface of medical-grade catheters, and demonstrate long term activity to prevent P. aeruginosa biofilm growth. Herein, we show that PslG h surface-modified catheters of polyethylene (PE), PDMS and polyurethane (PU) not only exhibit significantly reduced rates of biofilm formation under flow culture in vitro but also in vivo using a rat catheter infection model.