Probing the Role of Mucin‐Bound Glycans in Bacterial Repulsion by Mucin Coatings

Microbial colonization of implanted medical devices in humans can lead to device failure and life‐threatening infections. One strategy to prevent this unwanted colonization is to coat devices with polymers that reduce bacterial attachment. This study investigates how mucins, a class of biopolymers found in mucus, can be used as surface coatings to prevent attachment of selected respiratory pathogens to polystyrene surfaces. Our data show that coatings of porcine gastric mucins or bovine submaxillary mucins reduce surface attachment by Streptococcus pneumoniae and Staphylococcus aureus, but not Pseudomonas aeruginosa. To elucidate how mucin coatings repel S. pneumoniae and S. aureus, the molecular components of mucins are examined. Our data suggest that mucin‐bound glycans are key structural contributors of mucin coatings and are necessary for the repulsive effects toward S. pneumoniae and S. aureus.


Introduction
Microbial colonization is a leading cause of medical device failure. About 50% of indwelling devices become colonized by microbes, [ 1 ] causing a signifi cant fraction of hospital acquired infections. [ 2 ] Bacterial colonization begins with cell attachment to the device surface. Attached cells proliferate and mature to form resilient matrix-encased communities called biofi lms. Once established, biofi lms are diffi cult to eradicate due to their resistance to antimicrobial treatments. Hence, there is a strong focus on developing new surfaces to prevent bacterial attachment.
This study explores the natural mucus barrier, and specifically its gel-forming mucin polymers, for strategies to prevent surface attachment. Mucus is the hydrated polymer network that lines all wet epithelia in the human body, including the respiratory, digestive, and reproductive tracts. Mucins are highly glycosylated polymers, which exist in secreted and cellsurface forms. [3][4][5][6] Both mucin types can protect the underlying mucosal epithelia from microbial infection. For example, the secreted gastric mucin MUC5AC can maintain the bacterium Pseudomonas aeruginosa [ 7 ] and the yeast Candida albicans [ 8 ] and commercial bovine submaxillary mucins. First, mucin coatings were created by exposing polystyrene surfaces to mucin solutions. Mucin adsorption to polystyrene, presumably driven by hydrophobic interactions between the surface and the mucin protein core, [ 14,15 ] was verifi ed using fl uorescence microscopy and quartz crystal microbalance with dissipation monitoring (QCM-D) analysis, which confi rmed that the mucins formed relatively homogeneous coatings ( Figure S1, Supporting Information). To qualitatively assess the capacity for bacterial repulsion, mucin-coated polystyrene microtiter wells were exposed to bacteria. Attached bacteria were fl uorescently stained with SYTO 9 (Life Technologies), then visualized microscopically. Figure 1 A shows that both gastric and submaxillary mucin coatings reduced S. pneumoniae and S. aureus surface attachment compared to uncoated polystyrene surfaces, but had no effect on P. aeruginosa attachment. To quantify attachment, bacteria bound to the mucin-coated or uncoated polystyrene microtiter wells were evaluated using the CyQuant Assay (Life Technologies). Figure 1 B reveals that gastric mucin coatings reduced attachment of S. pneumoniae by 76.3% ± 8.6% and attachment of S. aureus by 81.3% ± 2.0% to the underlying polystyrene. Submaxillary mucin coatings were comparably effective, reducing attachment of S. pneumoniae by 71.5% ± 3.8% and of S. aureus by 81.0% ± 7.5% to the underlying polystyrene. Together, these data suggest mucin MUC5AC isolated from pig stomachs as a potential candidate biopolymer for the engineering of bacteriarepelling coatings. The data also extend our understanding of coatings generated by commercial submaxillary mucins to show that in addition to preventing S. aureus surface attachment, [ 12 ] they appear to repel S. pneumoniae . Importantly, these results highlight the limitations of mucins for universal bacterial-repelling surfaces because they appear ineffective against P. aeruginosa , a Gram-negative bacterium. This lack of repulsive effect is not generalizable to other Gram-negative bacteria because Escherichia coli , for example, can be repelled by mucin coatings. [ 13 ] Several mechanisms may contribute to the lack of effect toward P. aeruginosa . For example, when adsorbed to a surface, mucins may lose part of the biofi lm-suppressing functionality, which is exhibited in a 3D hydrogel network. Mucin-digesting enzymes secreted by P. aeruginosa [ 16 ] may also damage coating integrity and hence, its repulsive properties. Moreover, adhesins on bacterial surfaces, [ 17,18 ] which bind mucin-associated glycans [ 19,20 ] and mucin-peptide moieties, [ 21 ] may mediate interactions with the coatings.

Mucin-Bound Glycans Contribute to Repulsion of S. pneumoniae and S. aureus
Mucin-bound glycans within mucin coatings can contribute to the repulsion of mammalian cells. [ 22 ] To examine if mucinbound glycans also play a role in bacterial repulsion, bacterial  (3 of 6) 1500179 wileyonlinelibrary.com attachment was tested on coatings made from deglycosylated mucins, which are henceforth referred to as apo-mucins. Apo-mucins were generated from native mucins by chemical removal of mucin-associated glycans. Mucin deglycosylation was verifi ed using the periodic acid-Schiff (PAS) assay ( Figure 2 A). Apo-mucins were adsorbed to polystyrene surfaces to produce coatings, and fl uorescence microscopy of Alexa488labeled apo-mucin coatings confi rmed relatively homogeneous surface coverage ( Figure S2, Supporting Information). Figure 2 B shows that the removal of glycans from mucins reduced the coatings' ability to repel S. pneumoniae and S. aureus , but not P. aeruginosa , whose attachment was comparable on native mucin and apo-mucin coatings. Specifi cally, apo-gastric mucin coatings exhibited a 4.2-fold increase in S. pneumoniae attachment and a 10.8-fold increase in S. aureus attachment relative to native gastric mucin coatings (Figure 2 B). Similarly, apo-submaxillary mucin coatings had 3.1-fold more S. pneumoniae and 8.3-fold more S. aureus attached compared to their glycosylated counterparts (Figure 2 C). These data indicate that mucin-bound glycans contribute to the bacterial repulsion observed with gastric and submaxillary mucin coatings. The glycan compositions of gastric and submaxillary mucins differ considerably, [ 23,24 ] suggesting that the repulsive effect of the coatings observed here is dictated not by the specifi c glycan components, but instead by general physico-chemical properties conserved among the different mucin types.

Physico-chemical Analysis of Mucin Coatings
Deglycosylation alters the biochemistry and structure of the mucin polymers, which in turn will affect the overall properties of the mucin-coated surfaces. Generally, both the charge and hydrophilic properties of surfaces can affect their interactions with bacteria. [25][26][27][28][29] Therefore, these two properties were evaluated for the bacteria and the mucin-coated surfaces. First, we investigated the role of surface charge using zeta potential measurements. The data indicate that the surfaces of all three bacterial species used in this study had a negative zeta potential (Table S1, Supporting Information). To measure the zeta potential of mucin-coated surfaces, we used polystyrene beads (800 nm diameter) to which native or deglycosylated mucins were adsorbed. As depicted in Table 1 , uncoated polystyrene beads exhibited a stronger negative charge than beads coated with the fully glycosylated mucins. For comparison, apo-gastric mucins appeared to render the polystyrene surface more negatively charged than fully glycosylated native gastric mucins. Beads coated with apo-submaxillary mucins displayed no substantial change compared to glycosylated mucins (Table 1 ). One limitation of this experiment is that we were not able to identify if the changes in surface charge between native and apomucins stem from differences in surface adsorption, or from differences in the biochemical properties between the different mucin species. However, what may be concluded is that there is no measurable correlation between the zeta potential of the mucin-coated surfaces and their strength of repulsion, suggesting that surface charge is probably not the main parameter in this system to control bacterial surface adhesion.
In the next set of experiments, we used water contact angle measurements to investigate the hydrophilicity of the bacteria and the mucin-coated surfaces. Contact angles of bacteriacoated polystyrene slides revealed that all three species tested in this study were hydrophilic (Table S1, Supporting Information). Using polystyrene slides coated with the different mucin species, we showed that mucin-coated polystyrene surfaces were more hydrophilic than uncoated polystyrene (Table 1 ). This observation is consistent with previous fi ndings that mucins reduce the contact angle of polystyrene surfaces. [ 12 ] Our data also show that the deglycosylation of mucins did not substantially alter the contact angle of the mucin-coated surfaces. The lack of correlation between the contact angle of the different surfaces and their ability to prevent bacterial adhesion suggests that the hydrophilicity of the surface may not be the dominant factor to control microbial adhesion in this system. However, we note that these data need to be interpreted with caution, given their high variability.
As a third line of characterization, QCM-D was used to examine the hydrated thickness and softness of native mucin and apo-mucin coatings. Figure 3 A shows that gastric mucin coatings were 35.0 ± 9.9 nm thick, while submaxillary mucin coatings were 60.3 ± 3.2 nm thick. In contrast, apo-gastric mucins and apo-submaxillary mucins both formed thinner coatings that were less than 4 nm thick. QCM-D analysis also provided information about the softness of the coatings as a measure of energy dissipation of the acoustic waves. Figure 3 B shows that the dissipation was greater for coatings of native mucins than for their apo-mucin counterparts. Upon deglycosylation, the dissipation measurements decreased 21-fold for gastric mucin coatings and 2.5-fold for bovine submaxillary mucin coatings, indicating that apo-mucin coatings are stiffer than the native mucin coatings. Together, these data reveal that mucin-bound glycans provide mucin coatings with a certain thickness and softness, which are lacking in coatings generated with the apo-mucins (Figure 3 C ) . Thickness and softness can increase bacterial repulsion of coatings made from certain classes of synthetic polymers, such as polyethylene glycol. [30][31][32] Hence, also in the context of mucins, these parameters are likely important to modulate interactions with the bacteria. How the thickness and softness of coatings contribute to bacterial repulsion, whether by steric repulsion, the degree of hydration, or by other mechanisms, remains to be determined.
Adv. Mater. Interfaces 2015, 2, 1500179 www.advmatinterfaces.de www.MaterialsViews.com Table 1. Biochemical properties of mucin coatings from glycosylated and deglycosylated mucins. Zeta potential measurements were used to determine relative charge of native mucin and apo-mucin coatings on polystyrene beads. Contact angle measurements were used to determine hydrophobicity of native mucin and apo-mucin coatings. Reported values are mean ± standard deviation.

Conclusions
In humans, implanted medical devices are frequently a cause of nosocomial infections, which, if left untreated, can lead to high rates of device failure and severe systemic infection. In the body, mucins help protect the wet epithelia from microbial attachment and subsequent infection, and hence, show potential as building blocks for microbe-repelling surface coatings. This study shows that coatings of gastric mucins and submaxillary mucins on polystyrene surfaces prevent attachment of S. aureus and S. pneumoniae , two common respiratory pathogens. Our work furthermore suggests that mucin-bound glycans are necessary to prevent bacterial attachment, potentially by introducing a critical thickness and softness to the mucin coatings, which in the context of other polymer systems, is associated with enhanced antifouling properties. [30][31][32] Together, these fi ndings support a role for mucin-bound glycans in regulating host-microbe interactions. Mucin coatings could also inform the development of new antifouling materials that reduce unwanted microbial colonization of implanted medical devices.

Experimental Section
Preparation of Mucins : Bovine submaxillary mucins (Sigma-Aldrich) were dialyzed for 4 d against ultrapure water using a Spectra/Por Float-A-Lyzer G2 dialysis membrane (100 kDa M w cutoff, Spectrum Labs), then lyophilized for storage. Native porcine gastric mucins were purifi ed as previously reported. [ 33 ] Briefl y, mucus was scraped from fresh pig stomachs and solubilized in saline buffer with protease inhibitors and sodium azide. Insoluble material was pelleted by ultracentrifugation and mucins were purifi ed using size exclusion chromatography on a Sepharose CL-2B column. Mucin fractions were desalted, concentrated, and lyophilized for storage. To produce fl uorescent mucins, gastric and submaxillary mucins were labeled with Alexa488 (Life Technologies) following the manufacturer's instructions. Briefl y, Alexa488 succinimidyl ester in DMSO (10 µL, 10 mg mL −1 ) was added to mucins (1 mL, 3 mg mL −1 ) in bicarbonate buffer (0.2 M , pH 8). After incubation at room temperature for 1 h, free dye was separated from the labeled mucins using a Macrosep centrifugal fi lter (100 kDa M w cutoff, Pall).
Preparation of Apo-mucins : Mucins were deglycosylated by treatment with trifl uoromethanesulfonic acid (TFMS), followed by oxidation and beta-elimination of the residual sugars as previously described. [ 34 ] Lyophilized mucins (5 mg) were cooled on ice and mixed with icecold TFMS (375 µL) containing anisole (10% v/v). The solution was gently stirred on ice for 2 h then neutralized by the addition of a solution containing 3 parts pyridine, 1 part methanol, and 1 part water. Precipitates were dissolved by adding water. The solution was dialyzed for 2 d against ultrapure water using a dialysis membrane (20 kDa M w cutoff, Spectrum Labs). NaCl and acetic acid were then added to the solution (fi nal concentration of 0.33 M and 0.1 M , respectively). The solution was adjusted with NaOH to pH 4.5. For the oxidation step, icecold NaIO 4 (0.2 M ) was added to the mucin solution (fi nal concentration of 0.1 M NaIO 4 ), and incubated at 4 °C for 5 h in the dark. The unreacted periodate was destroyed by adding ½ volume of neutralizing solution containing Na 2 S 2 O 3 (0.4 M ), NaI (0.1 M ), and NaHCO 3 (0.1 M ). For elimination, the mucin solution was adjusted to pH 10.5 using NaOH (1 M ) and incubated at 4 °C for 1 h. The solution was then dialyzed overnight at 4 °C against NaHCO 3 buffer (5 × 10 −3 M , pH 10.5) and further dialyzed for 2 d against ultrapure water. The resulting apomucin solution was then concentrated and dissolved in the appropriate buffer.
Apo-mucins were evaluated for glycan removal using the periodic acid-Schiff (PAS) assay as previously described. [ 35 ] Briefl y, in a 96-well plate, a mixture (120 µL) of acetic acid (7%) and periodic acid (0.06%) was added to the sample (20 µL). The solution was incubated for 1.5 h at 37 °C. Schiff reagent (100 µL) was added to the wells and allowed to react at room temperature for 10 min before measuring absorbance (550 nm) using a SpectraMax M2 Microplate Reader (Molecular Devices). The ratio of absorbance to mucin mass contained in each sample was reported.
Microscopy of Mucin Coatings : To verify homogeneity, mucin coatings were visualized in the polystyrene microtiter wells. The wells were coated with gastric mucins, submaxillary mucins, or their apo-mucin counterparts labeled with Alexa488. A scratch was made in the fi lm with a pipette tip for reference. The fl uorescent coatings were imaged using an Axio Observer Z1 inverted epifl uorescence microscope (Zeiss) with a 40×/0.75 NA objective (Zeiss) and Hamamatsu OrcaR2 camera.
Measuring Bacterial Attachment to Mucin Coatings : Bacterial attachment to mucin coatings and apo-mucin coatings was evaluated for S. pneumoniae , S. aureus , and P. aeruginosa . Microtiter wells were coated with mucins as described above, or left untreated as a control. S. pneumoniae TIGR4 serotype 19F was cultured in Todd Hewitt Broth (Becton Dickinson) supplemented with 0.5% yeast extract (Becton Dickinson) in static conditions at 37 °C with 5% CO 2 . S. aureus UAMS-1 was cultured in Brain Heart Infusion (Becton Dickinson). P. aeruginosa Adv. Mater. Interfaces 2015, 2, 1500179 www.advmatinterfaces.de www.MaterialsViews.com  Zeta Potential : Zeta potential measurements for mucin coatings were obtained by adsorbing native mucins or apo-mucins on the surfaces of polystyrene beads (≈800 nm diameter, Sigma-Aldrich). Beads (4 µL of a 10% solids solution) were incubated for 1 h in native mucins (100 µL, 200 µg mL −1 ) or apo-mucins (100 µL, 200 µg mL −1 ) in HEPES (0.02 M , pH 7.4), then washed 3 times with PBS. Zeta potential measurements were carried out in PBS using a ZetaPALS Zeta Potential Analyzer (Brookhaven Instruments). As a control, zeta potential was measured for uncoated polystyrene beads in PBS. Zeta potentials of bacteria were measured using a solution of bacteria OD 600 0.05 in PBS. Measurements were performed in triplicate.
Contact Angle : Contact angle measurements for mucin coatings were obtained using native mucins or apo-mucins adsorbed on polystyrene slides (Electron Microscopy Sciences). Native mucins (100 µL, 200 µg mL −1 ) or apo-mucins (100 µL, 200 µg mL −1 ) in PBS were incubated on polystyrene surfaces for 1 h at room temperature. The coatings were washed with PBS 3 times, followed by a dH 2 O wash to prevent salt crystal formation. Contact angle measurements for bacterial surfaces were obtained using bacteria adsorbed to polystyrene slides. Bacteria (100 µL) at OD 600 0.4 in PBS were incubated on polystyrene slides for 3 h, followed by a wash with MilliQ water. The mucin or bacteria coatings were then air dried for 1 h, and the contact angle of a MilliQ water drop (≈5 µL) was measured using a goniometer (cameraequipped VCA 2000, AST Products). The reported values are averages of triplicate measurements of the advancing angle, meaning the constant angle between the liquid and the surface as the drop increased in volume.
QCM-D : Quartz crystal microbalance with dissipation monitoring (QCM-D, E4 system, Q-Sense) was used to measure the hydrated mass of mucins adsorbed to a polystyrene-coated quartz crystal (QSX305, Q-sense). Solutions of native mucins (200 µg mL −1 ) or apo-mucins (200 µg mL −1 ) were adsorbed to the crystal. The crystal vibration was followed at its fundamental frequency (≈5 MHz) and fi ve overtones (15,25,35,45,55, and 65 MHz). Once the excitation was stopped, changes in the resonance frequencies and dissipation of the vibration were followed at the six frequencies. When the adsorbed layers are highly hydrated, they usually possess viscoelastic properties requiring the measurement data to be modeled. Hydrated thickness was calculated using the Q-Tools 3.0.12.518 software that includes the Voigt model (i.e., a spring and dashpot in parallel under no slip conditions), [ 36 ] assuming a density of 1050 kg m −3 (as validated for a related system) [ 37 ] and that the coating is homogeneous in thickness and over the crystal's surface. Measurements were performed at least in duplicate.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.