Surface-associated MUC5B mucins promote protease activity in Lactobacillus fermentum biofilms

In order to investigate the naturally occurring interactions between MUC5B and Lactobacillus fermentum, a clinical strain known to be able to survive and grow in a mucus-rich environment was isolated from dental plaque. The strain, which originated from approximal supra-gingival plaque from a healthy 30-year old male, was identified by selective growth in Rogosa medium and fermentation tests using the API-50 CHL system (BioMerieux, Marcy l’Etoile, France). Identification as L. fermentum was confirmed by sequencing of the pheS gene[19]. Bacteria were stored at −80°Cand sub-cultured twice on blood agar in an atmosphere of 5% CO₂ in air at 37°C for 48 hours before use.

Human salivary mucin MUC5B was isolated as previously described[20]. Whole saliva was collected on ice from eight volunteers, pooled and then mixed with an equal volume of 0.2 M NaCl and stirred overnight at 4°C. After gentle centrifugation (4400 g for 30 min at 4°C), the supernatant was subjected to isopycnic density-gradient centrifugation in CsCl/0.1 M NaCl (Beckman Optima LE-80 K, rotor 50.2Ti, 36,000 r.p.m., 96 h, 15°C, start density 1.45 g/ml; Beckman, Fullerton, CA, USA) and 22 fractions were collected. The MUC5B-containing fractions were pooled, dialysed against PBS (0.15 M NaCl, 5 mM NaH₂PO₄, pH 7) and stored at −20°C until use. To determine the concentration of MUC5B, aliquots of the MUC5B-containing fractions were pooled, dialyzed against water, lyophilized, and weighed (0.3 mg ml^-1). The pool of MUC5B was subjected to sodium dodecyl sulfate–poly-acrylamide gel electrophoresis (SDS–PAGE), to examine whether low-molecular-weight proteins associated with the gel network contaminated the preparation.

L. fermentum colonies from blood agar were suspended in nutrient broth [Todd-Hewitt broth (Difco Laboratories, Detroit, MI, USA)] or in a MUC5B solution in PBS to give OD₆₀₀ values of 0.4±0.01 (corresponding to approximately 1×10⁶ cells ml^–1). Biofilms were established in mini-flow cell Ibidi μ-slides (Integrated Bio Diagnostics, Martinsried, Germany) by inoculating 120 μl of the suspension into the channels and incubating in humidified 5% CO₂ in air at 37°C for 24 hours. The channels were then rinsed with PBS to remove non-adherent cells. In some cases, the slides were pre-conditioned with 100 μl of MUC5B+10 μl 10 mM CaCl₂ overnight in humidified 5% CO₂ in air at 37°C to obtain a MUC5B coating. The slides were rinsed with PBS prior to biofilm formation. Biofilm formation was confirmed by staining with the BacLight™ LIVE/DEAD kit (Molecular Probes Inc., OR, USA) and visualization using a Nikon Eclipse TE2000 inverted confocal scanning laser microscope (CSLM).

To investigate the proteolytic activity of planktonic cells, L. fermentum cells were examined with the fluorescent protease assay kit QuantiCleave™ (Pierce, Rockford, IL, USA) as described previously[21]. Briefly, cells were incubated for 1 hour at 37°C with the FITC labeled casein substrate at a concentration of 10 μg ml^-1 in 25 mM Tris with 150 mM NaCl and adjusted to pH 7.2. Aliquots were then introduced into mini-flow-cell Ibidi μ-slides and counterstained with the red fluorescent DNA-staining dye SYTO62® (Molecular Probes Inc., OR, USA) for 10 minutes at room temperature. For biofilm cells, the fluorescent substrate and counterstain were introduced directly into the mini-flow cells in which the biofilms were growing. Cells were then visualized using CSLM and 20 random images from each biofilm slide were captured using a motorized station connected to the Nikon software of the microscope. Each image contained more than 1000 bacterial cells and the proportion of red and green cells for each biofilm slide, corresponding to approximately 20,000 cells, was analysed using the software package bio Image_L[22]. Experiments were repeated three times and the mean percentage value for proteolytically active cells was calculated. Results were then further analyzed using the Mann–Whitney U test and p values of less than 5% were regarded as significant.

Mini flow-chambers were pre-conditioned with MUC5B and L. fermentum cells then inoculated in nutrient broth. After 24 hours, the medium was removed and MUC5B mucins suspended in PBS were added to the biofilm for an additional 24 hours. After this time, the fluid in the flow-cell was collected with a pipette and centrifuged to remove bacterial cells (10000 g, 5 min, 4°C). The mucin-containing supernatant, as well as a control (MUC5B mucins incubated for 24 hours at 37°C but not exposed to L. fermentum) was analysed using SDS-PAGE under denaturing conditions followed by Western blotting. Samples (15 μl of the supernatant) were mixed with 5 μl NuPAGE LDS sample buffer (Invitrogen, Carlsbad, CA, USA) and run on NuPAGE 4–12% BisTris gels (Invitrogen, Carlsbad, CA, USA) at 180 V for 1 hour. The gels were electro-blotted onto PVDF membranes (Millipore, Immobilon-P, 0.45 mm, Bedford, MA, USA) overnight using a Mini Trans-Blot Electrophoretic Transfer Cell (Bio-Rad, Hercules, CA, USA) and the membranes then blocked with 5% (w/v) dry milk in TBST (20 mM Tris, 137 mM NaCl, 0.05% Tween 20) for 1 hour. MUC5B mucins were detected using an antiserum (LUM5B-14), raised against a synthetic peptide with the sequence CRAENYPEVSIDQVGQVL present in the third Cys domain of the MUC5B peptide, diluted 1:1000 in 5% BSA/TBST. The membranes were then incubated (1 hour) with a horseradish peroxidase-conjugated goat anti-rabbit antibody (Dako, P0448, Glostrup, Denmark) diluted 1:2500 in 5% skim milk/TBST, and bands were visualized using the ECL Western detection kit (Pierce, Rockford, IL, USA).

The bulk phase surrounding biofilm cells growing in a MUC5B- or a nutrient broth- environment, was removed with a pipette from the flow-cells and replaced with 50 μl of extraction buffer consisting of 5 M urea, 2 M thiourea, 2% CHAPS, 2% caprylyl sulfobetaine, 2 mM tributyl phosphine, 0.5-2% IPG and 40 mM Tris base adjusted to pH 9.5. In order to solubilize surface proteins from the remaining biofilm cells, the flow-cells were incubated with shaking for 1 hour at room temperature and the extraction buffer then removed and centrifuged for 10 minutes (6000g). The protein concentrations in the resulting supernatants were determined using a 2D Quant kit (GE Healthcare Life Sciences, Little Chalfont, UK). The supernatants were then diluted with rehydration buffer containing 8 M urea, 2% CHAPS, 10 mM DTT, 2% immobiline (IPG) buffer (GE Healthcare Life Sciences, Little Chalfont, UK) to give a final protein concentration of 20 μg in 200 μl. These samples were placed in re-swelling cassettes with 11 cm IPG strips, pH range 4–7, on top and rehydration was performed under oil at room temperature for 30 hours. Isoelectric focussing was carried out essentially as described by Davies et al.[23], using the Multiphor II (Amersham Pharmacia Biotech, Little Chalfont, UK), for 85, 500 volt hours. The strips were equilibrated and then embedded on top of 14% polyacrylamide gels. SDS-PAGE was performed at a constant current of 15 mA/gel, 10°C, over-night in a PROTEAN II xi cell (Bio-Rad, Hercules, CA) with high-range Rainbow molecular-weight standards (GE Healthcare Life Sciences, Little Chalfont, UK) on the acidic side of the IPG strips. Gels were stained with silver according to the manufacturer’s instructions or colloidal Coomassie Brilliant Blue G for protein identification. Gels were analyzed using Delta2D software (Decodon GmbH, Greifswald, Germany). After identification of the spot boundaries, the integrated optical density (IOD) values were determined and expressed as a percentage of the total intensity per gel. Spots showing more than a three-fold difference were cut from the Coomassie-stained gels, subjected to in-gel tryptic digestion and analysed by LC-MS/MS as described[23]. For detection of O- sialoglycoprotein endopeptidase, 2DE gels were electro-blotted onto PVDF membranes (Millipore, Immobilon-P, 0.45 mm, Bedford, MA, USA) overnight using a Mini Trans-Blot Electrophoretic Transfer Cell (Bio-Rad, Hercules, CA, USA) and the membranes then blocked with 5% (w/v) dry milk in TBST (20 mM Tris, 137 mM NaCl, 0.05% Tween 20) for 1 hour. The protein was detected using a polyclonal antiserum MOB-LFGE-2 raised against a synthetic peptide with the sequence (CDAAGEAYDKVGRV) (Innovagen AB, Lund, Sweden), diluted 1:1000 in 5% BSA/TBST. The membranes were then incubated (1 hour) with a horseradish peroxidase-conjugated goat anti-rabbit antibody and bands were visualized using the ECL Western detection kit as described above.

To investigate the effect of MUC5B on proteolytic activity in planktonic cells of L. fermentum, cells were exposed to MUC5B in the fluid phase. Nutrient broth was used as a control. The proportion of proteolytically-active cells was evaluated using CLSM after incubation with FITC-casein. This revealed no difference between the proportion of proteolytically-active cells in nutrient broth to those in MUC5B mucins suspended in PBS, and in both cases the activity in the population was lower than 4% (Figure 1). No difference in the level of activity of the cell population suspended in MUC5B mucins in PBS and that suspended in PBS alone (control) was seen (data not shown). To compare the proteolytic activity of planktonic cells with that of biofilm cells of L. fermentum, bacteria were allowed to form biofilms under static conditions in a mini-flow chamber system for 24 hours in a MUC5B environment (on surfaces coated with MUC5B and MUC5B in the fluid phase) or a nutrient broth environment. The proportion of proteolytically-active cells in biofilms in the MUC5B environment was ten-times greater (47±0.6%) than that for their planktonic counterparts (Figure 1b) whereas no such increase was seen in biofilms formed in the presence of nutrient broth. Imaging of the biofilms showed that L. fermentum cells adhered to both the uncoated and the MUC5B-coated surfaces and formed biofilms (Figure 2). In the MUC5B environment, proteolytically-active cells were evenly distributed throughout the biofilm (Figure 2a) whereas activity in the nutrient broth environment was low (Figure 2b). These data indicate that proteolytic activity is enhanced in biofilm cells of L. fermentum in a mucin-rich environment. The effect is not seen when planktonic L. fermentum cells are exposed to the same mucins in the fluid phase or in the presence of nutrient broth.

To determine whether the enhancement of proteolytic activity in the L. fermentum biofilm population in the MUC5B environment was mainly attributable to the surface-associated mucins or those in the bulk phase, the effects of these were investigated separately. Biofilms of L. fermentum were therefore either established on MUC5B-coated surfaces with nutrient broth in the fluid phase, or on uncoated surfaces with MUC5B in the fluid phase. In the presence of surface-associated MUC5B, image analysis revealed that 13±3% of the population was proteolytically active (Figure 3) while in cells exposed to MUC5B in the fluid phase only, activity was very low (0.3±0.02%). Thus, these data show that the presence of mucins on the surface was important for the enhancement of proteolytic activity. However, the proportion of active cells on the surface-associated mucins alone was significantly lower than that seen in the MUC5B environment, where mucins were also present in the bulk phase. This suggests that adherence to surface-associated MUC5B may have ‘primed’ the cells allowing them to respond to the presence of MUC5B in the fluid phase.

To determine whether mucins in the mucin-rich environment are degraded by the proteases produced within the L. fermentum biofilms, MUC5B mucins which had been in contact with biofilms for 24 hours were subjected to SDS-PAGE followed by Western blotting with an anti-MUC5B antiserum (Figure 4). Control mucins, which had not been in contact with bacteria, were visualized as a single band on top of the gel showing that they were of high-molecular-weight (>300 kDa). However when mucins had been incubated with L. fermentum biofilms, an additional band appeared in the upper region of the 4-12% gradient gel indicating that part of the mucins had undergone some degradation that allowed MUC5B to migrate into the gel. Thus these data confirm that proteolytically-active biofilm populations of L. fermentum can degrade MUC5B.

Since MUC5B mucins were degraded by biofilm cells of L. fermentum, the genome database was scanned for proteases capable of degrading glycoproteins. Using this approach, one protease, O- sialoglycoprotein endopeptidase, with a molecular weight of approximately 35 kDa, was selected as a possible candidate. Western blotting of 2DE gels of surface proteins from L. fermentum biofilm cells with a polyclonal antiserum raised against a peptide in the sequence of the glycopeptidase identified a spot at 32 kDa, corresponding to O- sialoglycoprotein endopeptidase (see Figure 5a). To investigate whether this protein was differentially expressed on the surface of L. fermentum biofilm cells in the different environments, surface proteins from biofilm cells in a MUC5B environment were compared with those from cells in a nutrient-broth-environment using 2D-gel electrophoresis. Image analysis of silver stained gels revealed that expression of the O- sialoglycoprotein endopeptidase was enhanced 2.7-fold in the presence of MUC5B, as compared to nutrient broth. In addition, a number of spots in the range 52–76 kDa showed higher levels of expression in the MUC5B- than in nutrient-broth-environment (see Figure 6 and Table 1). These were cut from a corresponding Coomassie-stained gel (see Figure 5b), subjected to LC-MS/MS and the proteins then identified by comparison of the tryptic fragments with the Mascot database. This revealed four proteins with a more than three-fold up-regulation in the MUC5B- as compared to the nutrient-broth-environments, corresponding to the chaperone proteins; trigger factor (25-fold increase), DnaK (9-fold increase), Ef-G (6-fold increase) and GroEL (3-fold increase). Thus these data show that besides up-regulation of the glycoprotease, increased proteolytic activity in the population is associated with an increase in the surface-expression of the chaperone proteins DnaK, GroEL and trigger factor.