Abstract
Background To defend the lungs, mucus adheres to bacterial cells and facilitates their removal by ciliary transport. Our goals were to measure the affinity of mucus for the respiratory pathogen Staphylococcus aureus and identify bacterial genes that regulate this interaction.
Methods S. aureus was added to pig tracheas to determine whether it binds mucus or epithelial cells. To quantify its affinity for mucus, we developed a competition assay in microtiter plates. Mucin was added over a dose range as an inhibitor of bacterial attachment. We then examined how transcriptional regulator MgrA and cell wall transpeptidase sortase (SrtA) affect bacterial interaction with mucin.
Results In pig tracheas, S. aureus bound mucus strands from submucosal glands more than epithelial cells. In microtiter plate assays, ΔsrtA failed to attach even in the absence of mucin. Mucin blocked wild type S. aureus attachment in a dose-dependent manner. Higher concentrations were needed to inhibit binding of ΔmgrA. Co-deletion of ebh and sraP, which encode surface proteins repressed by MgrA, suppressed the ΔmgrA binding phenotype. No differences between ΔmgrA and wild type were observed when methylcellulose or heparin sulfate were substituted for mucin, indicating specificity.
Conclusions Mucin decreases attachment of S. aureus to plastic, consistent with its physiologic role in host defense. S. aureus deficient in MgrA has decreased affinity for mucin. Ebh and SraP, which are normally repressed by MgrA, may function as inhibitors of attachment to mucin. These data show that specific bacterial factors may regulate the interaction of S. aureus with mucus.
Introduction
Mucus is an important defense against respiratory infections. It is a gel-like material comprised of hydrated mucin proteins and inorganic salts produced on epithelial surfaces. In large animals, airway mucus is produced in the form of strands arising from submucosal glands (1-3). Inhaled particles and pathogens can become entrapped in mucus. After the release of mucus strands from submucosal gland ducts or goblet cells, ciliated cells transport mucus together with any particles trapped in the mucus, expelling them from the airway (1). Although mucus is colloquially described as “sticky,” its adhesive properties are seldom quantified. Proper physiologic function of mucus would require that it have some affinity for a diverse repertoire of particles and pathogens, but not excessively adhere to the epithelium or gland duct surfaces. In cystic fibrosis (CF), mucus may interact with pathogens, but it fails detach from gland ducts and is unable to be carried away by ciliary beating. This results in impaired transport of mucus out of the airway (3). In CF pigs, mucus strands fail to detach from airway submucosal glands and becomes pathologically immobile. Thus, it has been hypothesized that mucus stasis may contribute to the establishment of the chronic bacterial infections in CF.
S. aureus is a typical pathogen in CF, as approximately 70% of people with CF are infected with the bacteria (4, 5). Among CF pathogens, S. aureus has the highest infection incidence in the preschool age group, making it one of the earliest organisms in CF (4). Furthermore, patients infected with methicillin-resistant S. aureus (MRSA) experience faster CF disease progression (6-8). Currently there is no vaccine against S. aureus, and eradication of incident infections (especially MRSA) remains challenging (9). Hence, there is a critical need to understand how S. aureus initially infects the CF airway.
There is some prior evidence the S. aureus associates with mucus. Mucins from saliva, nasopharynx, and conjunctiva are known to bind S. aureus (10). In animals infected intranasally with S. aureus, bacteria bind to mucus from anterior and posterior turbinates without direct binding to the underlying epithelium (11). However, the molecular basis for S. aureus attachment to mucin is not fully understood. S. aureus expresses a variety of surface proteins, collectively known as adhesins, that are implicated in the pathogenesis of other serious infections like endocarditis and sepsis. We hypothesized that bacterial genetic factors regulating adhesins control S. aureus attachment to mucus. In this study, we observed whether S. aureus attaches to mucus in pig tracheas and we developed a method to quantify the affinity of S. aureus for mucin by a competition assay. We used this approach to measure the interaction of mucin with S. aureus strains with suspected defects in attachment. We examined two candidate regulators of bacterial attachment: Sortase A (SrtA), a transpeptidase enzyme required for surface display of proteins (12), and multiple gene regulator A (MgrA), a transcriptional regulator of surface adhesins, including the giant staphylococcal surface proteins Ebh and SraP (13-15).
Materials and methods
Bacteria strains and growth conditions
Bacterial strains and plasmids used are listed in the Table 1. S. aureus was grown overnight in tryptic soy broth (TSB), supplemented with 10 µg/ml of chloramphenicol (Cam) to maintain the sGFP-expressing plasmid pCM29. For construction of S. aureus mutants, tryptic soy agar with Cam or with tetracycline (Tet, 1 µg/ml) were also used. To construct mutant strains, bacteriophage transduction between S. aureus strains was performed with phage 80α or 11 as described previously (16). To generate the sortase mutant AH4480, the ΔsrtA mutation containing a TetR cassette was transduced from the previously described ΔsrtA mutant in the S. aureus strain RN4220.
Trachea preparations
Animal studies were approved by the University of Iowa Animal Care and Use Committee. Male and female newborn pigs were obtained from Exemplar Genetics. The pigs were sedated with ketamine and acepromazine or xylazine and euthanized with I.V. Euthasol. We removed 1 cm tracheal segments following euthanasia, made a longitudinal cut on the ventral surface, and pinned the trachea flat to expose the mucosal surface. To mimic the CF electrolyte transport defect, we used conditions that impair chloride and bicarbonate transport. We submerged the tissue with 40 ml HEPES buffered saline with 10 µM bumetanide as previously described (3). 100 µM methacholine was added to stimulate submucosal gland secretion (1, 3).
Imaging interaction of S. aureus and mucus in pig tracheas
S. aureus strains were grown overnight in tryptic soy broth (TSB) with 10µg/ml Cam to maintain selection of GFP plasmids (Table 1). We added 40 µL of stationary phase GFP-S. aureus to the 40 ml HEPES buffered saline. In some experiments, we added 1:10,000 red fluorescent nanospheres (FluoSpheres, Molecular Probes) to identify mucus strands (1, 3). We used a 25x (Nikon Apo LWD, Water, NA 1.1) objective and a Nikon A1R confocal microscope. We quantitated the GFP signal on the epithelial surface and compared to the GFP signal on mucus strands. We determined the speed of individual S. aureus cells on these pig tracheas using Imaris software (Bitplane) and compared the speed of free bacteria to attached bacteria.
Quantifying the interaction of S. aureus and mucus by competition assay
Pig Gastric Mucin (type III, sialic acid 0.5-1.5%, Sigma #M1778, Lot SLBL7748V) was dissolved in phosphate buffered saline, pH 7. Mucin concentrations ranging from 4 µg/mL to 8 mg/mL were produced by serial dilution. 100 µL of diluted mucin was added to non-treated 96-well black optical plates (Thermo Fisher #265301). Methylcellulose (MP Biomedicals #155492, Lot 6257E), heparin sulfate (Sigma #H3149-250KU, Lot SLBS9674), and bovine submaxillary mucin (Sigma #M3895-1G, Lot SLBS0651V) were obtained commercially and prepared similarly as antagonists of bacterial attachment. To dissolve methylcellulose, we autoclaved it as a dry powder and dissolved it in PBS by stirring overnight at 4 °C.
For in vitro binding assays, we diluted stationary-phase GFP S. aureus to OD of 0.6 in PBS. Each strain of S. aureus had similar GFP signal per cell. 50 µL of S. aureus were added per well following dilution. Bacteria were allowed to attach for 1 hour at 37 °C. The mucin suspension and any bacteria not attached to the plastic surface were removed by aspiration. We performed two rounds of washing with PBS followed by suctioning. The residual fluorescence was assayed using a SpectraMaxi3X plate reader with settings optimized for GFP.
Statistical analysis
For each experiment, we used 4 replicate wells for each condition. Following experiments, we calculated the average GFP fluorescence from replicate wells and plotted the average fluorescence versus log10 of the final concentration of binding antagonist measured in mg/mL. We used GraphPad Prism version 7 to perform 4-parameter logistic regression, allowing for calculation of IC50. We rejected experiments from further analysis if there were not at least 2 top and bottom values for fitting the logistic regression. For summary statistics, we used IC50 values from independent experiments performed on different days. For comparison of IC50 values, we used log-transformed data for statistical tests. For non-normally distributed data sets, we used non-parametric statistical tests as appropriate.
Results
S. aureus attaches to mucus strands in pig tracheas
In a previous study, we observed that mucus strands fail to detach after they emerge from submucosal glands in CF pig tracheas (3). If bacteria attach to these mucus strands, they may be retained in the airway. We obtained tracheas from newborn wild type pigs, submerged them in bicarbonate-free HEPES buffered saline and treated with bumetanide. These conditions block chloride and bicarbonate secretion, simulating CF airway conditions (3). We added methacholine to stimulate mucus secretion by submucosal glands. To these preparations, we added stationary phase S. aureus expressing GFP to test whether these bacteria associate with mucus strands, or if they bind directly to the epithelium (Figure 1). We selected LAC, a community-acquired strain of MRSA (CA-MRSA), as our model organism. CA-MRSA has recently increased in prevalence within the CF population (17).
To help identify mucus strands, we added red fluorescent nanospheres as previously described (1, 3). In the presence or absence of nanospheres, GFP S. aureus was predominantly located on mucus strands (Figure 1A-C). Under these experimental conditions, mucus strands move slowly and do not properly release from submucosal gland ducts (3). S. aureus attached to these mucus strands had average speed less than 1 mm/min, whereas S. aureus that were not attached to strands moved more quickly, indicating that few were attached to the surface epithelium (Figure 1D, Supplemental Videos 1 and 2).
Quantifying bacteria-mucus interaction by competition assay
Mucus strands form a discontinuous mesh rather than a complete gel over the airway surface (1, 2). Such a discontinuous layer may not prevent bacteria from reaching the epithelial surface. However, our confocal microscopy observations indicate that when S. aureus are added to pig tracheas, they accumulate on these mucus strands rather than on the epithelial surface. This might indicate that mucus strands function as competitors with the epithelial surface for bacterial binding, similar to a decoy receptor. Once bacteria attach to mucus strands, ciliary beating transports the aggregate of mucus and bacteria from the lungs. To quantitate the adherent function of mucus, we designed a competition assay to measure how efficiently standardized mucins prevent the attachment of bacteria to a standardized surface without the confounding effect of ciliary flow.
We used microtiter plates as a standardized surface for attachment. GFP-labeled S. aureus were added to these plates in the presence of variable concentrations of pig gastric mucin. We incubated the suspension of bacteria and mucin for one hour at 37 °C, allowing bacteria to associate with either the mucin or the plastic surface. Bacteria that did not attach to the plastic surface, but instead associated with mucin in solution, were removed by aspiration and washing. We quantitated residual bacteria attached to the surface with a fluorescent plate reader (Figure 2).
We first compared the attachment of wild-type LAC and mutant strains to plastic in the absence of mucin. We compared LAC to an isogenic mutant lacking Sortase A (ΔsrtA) (12), which catalyzes attachment of S. aureus surface proteins to the cell wall, and MgrA (ΔmgrA), a transcriptional regulator of adhesin expression (13). All three strains had similar attachment to plastic prior to removal of unattached bacteria. ΔsrtA had significantly decreased GFP signal following washing and aspiration, indicating defective attachment to plastic (Figure 3A). We did not study it further. By contrast, the GFP signal from ΔmgrA was indistinguishable from LAC.
Mucin prevents bacterial attachment to plastic in a dose-dependent manner
When pig gastric mucin was added to wells with bacteria, there was a saturating dose-dependent reduction in attachment of wild-type LAC to plastic (Figure 3B). Using 4-parameter logistic regression, we determined that the IC50 of pig gastric mucin as an inhibitor of wild type S. aureus attachment to plastic was approximately 1.6 x 10−2 mg/mL.
MgrA is necessary for normal mucin interactions
Using this method, we compared LAC with ΔmgrA, a strain with significantly altered surface adhesin expression (13). Although both LAC and ΔmgrA had similar binding capacity for the microtiter dish surface in the absence of mucin (Figure 3A), we observed differences in absorption to the plastic surface in the presence of mucin (Figure 3B). Significantly higher concentrations of pig gastric mucin were required to block attachment of ΔmgrA, as visualized by the rightward shift in the inhibition curve. We repeated these studies using a partially complemented ΔmgrA strain (13) and found that partial restoration of MgrA partially normalized the interaction with mucin (Figure 3C).
MgrA-dependent differences in interaction with mucins, but not other biopolymers
To validate the observed genotype-dependent binding differences were specific to mucin, we compared adhesion of LAC and ΔmgrA to mucin versus relevant biologic polymers: methylcellulose, bovine submaxillary mucin, and heparin sulfate. Methylcellulose is used as a mucin substitute in pharmacologic products such as artificial tears (18). We compared it to mucin as both share viscoelastic physical properties and repeating saccharide units. Bovine submaxillary mucin is a secreted mucin with a notably higher sialic acid content than porcine gastric mucin (up to 30% vs. 1%, respectively) (19, 20, 21). We tested whether the increase in sialic acids displayed by BSM would lead to altered attachment by ΔmgrA, since SraP was previously shown to interact with sialic acid (22). Heparin sulfate is a linear polysaccharide abundantly secreted at mucosal surfaces. We compared it to mucin as it shares similar glycosylated residues but carries a stronger negative charge than mucin due to its sulfated moieties.
Figure 4 summarizes our results. We found the altered binding phenotype observed in the ΔmgrA mutant was specific to mucins. Both porcine gastric mucin and Bovine submaxillary mucin exhibited statistically significant IC50 differences between LAC (WT) and ΔmgrA strains. Conversely, methylcellulose and heparin sulfate failed to recapitulate bacterial genotypic differences in mucin adhesion.
These studies demonstrated that mucin is a poorer inhibitor of ΔmgrA attachment compared to the wild type strain LAC. We considered three possible explanations for this phenomenon. 1) MgrA is needed for the expression of classical S. aureus surface adhesins (13); 2) MgrA suppresses expression of surface proteins that block interaction with mucin; or 3) MgrA induces the expression of adhesins that directly bind mucin.
A S. aureus mutant lacking conventional adhesins has normal mucin interaction
Some of the most well studied adhesins in S. aureus are ClfA, ClfB, and the fibronectin binding proteins A and B (FnbpAB). These adhesins have broad roles in binding fibrinogen, fibronectin, and other extracellular matrix components, and have been linked to nasal colonization, clumping and general S. aureus virulence properties (23). Using a mutant defective in all these adhesins (LAC ΔclfAB ΔfbnAB), we tested interactions with mucin. We found that the inhibition curve of pig gastric mucin versus this clumping factor mutant was indistinguishable from wild-type LAC (Figure 5A). Thus, the presence or absence of these important adhesins does not explain the altered interaction of ΔmgrA with pig gastric mucin.
Deficiency in Ebh and SraP suppress ΔmgrA mucin interaction phenotype
MgrA functions as a transcriptional repressor that blocks the expression of the large surface proteins Ebh and SraP in the S. aureus LAC strain (13). We hypothesized that increased expression of one of these surface proteins was responsible for the decreased mucin binding observed in ΔmgrA. Therefore, we tested the interaction of mucin with ΔmgrA lacking either Ebh, SraP, or both (Figure 5 B and C). We found that deficiency in Ebh or SraP alone had minimal effect of mucin interaction compared to the parental ΔmgrA strain. However, the loss of both Ebh and SraP completely restored binding to the ΔmgrA strain. These data are most consistent with Ebh and SraP functioning as interchangeable factors that inhibit bacterial interaction with mucin. This is analogous to their function as anti-clumping factors (13).
Discussion
S. aureus respiratory infections remain a major problem for patients with CF. It is well documented that infection with methicillin-resistant strains of S. aureus is associated with worsening CF disease progression (6, 7), and newer treatments for CF have had minimal effect on the prevalence of S. aureus (24, 25). Moreover, S. aureus is a major cause of respiratory disease in children and adults without CF (26, 27, 28). Although S. aureus may colonize asymptomatic individuals (29), S. aureus is a prevalent cause of community-acquired pneumonia, especially following influenza infection (30-32) and has high case fatality rates (27, 33, 34).
S. aureus binds airway mucus
Because of its importance as a respiratory pathogen, it is crucial to understand how S. aureus initiates infections. We found that when introduced into explanted airways, S. aureus accumulates on strands of mucus arising from submucosal glands. When liquid secretion from these glands is blocked pharmacologically, the S. aureus that attach to these strands may not be cleared efficiently (3). Previous studies of S. aureus showed minimal contact with ciliated mucosal surface in human nasal biopsies (35) and that attachment could be decreased with lectin or monosaccharide competition (36), consistent with bacterial interaction with mucus.
Mucus is described as “sticky,” but its adhesive properties are seldom quantified. The development of a competition assay enabled us to measure the strength of the interaction between mucus and bacteria. Using the IC50 of competition curves as an estimate of the affinity of S. aureus for mucin, we found that the interaction with wild-type S. aureus is relatively weak (∼10−2 mg/mL). However, mucin production is abundant and thus it may have high capacity for removing inhaled bacteria. Since S. aureus can stimulate mucin synthesis (37, 38), initial infection could amplify the number of potential binding bacterial sites. This may perpetuate infection in CF, where the release of mucus is impaired (3).
Bacterial surface factors regulate attachment and mucus interaction
How does S. aureus adhere to respiratory mucus? Differences between bacterial strains isolated from patients suggest genetic factors control the interaction of S. aureus with mucus (39). We predicted that genes regulating the display of adhesins may alter the attachment properties of S. aureus (23). Most of the classical surface adhesins are added to the S. aureus cell wall by sortase-dependent transpeptidation (12). However, we observed a deficiency in attachment to our standardized surface with ΔsrtA S. aureus and did not study them further. MgrA regulates the expression of some adhesins on the bacterial cell wall (13, 15). In the absence of MgrA, higher concentrations of mucin were required to block attachment of S. aureus to microtiter plates, consistent with decreased interaction of S. aureus with mucin. The decreased interaction of ΔmgrA with mucin was not related to well-known adhesins ClfA, ClfB, or FnbpAB, as a different strain lacking these surface proteins had normal interaction with mucin. Instead, the binding phenotype of ΔmgrA depended on Ebh and SraP, which MgrA normally represses. We speculate that these large surface proteins sterically block the interaction of glycans from mucin side chains with low affinity receptors on the bacterial cell wall. This theory parallels recent work demonstrating decreased adhesion of ΔmgrA S. aureus to various host molecules, including fibrinogen, fibronectin, and collagen (15). This may also be a mechanism for the anti-clumping activity of these large bacterial surface proteins (13).
S. aureus interaction with mucin may weaken with greater negative charge
Previous studies of S. aureus showed its attachment to mucin was only weakly inhibited by N-acetylneuraminic acid (11), and could be augmented by neuraminidase treatment (40). This implies a low affinity state between S. aureus and sialic acid, likely explained by charge repulsion between teichoic acids on the bacterial cell wall and sialic acid residues on mucins. We found that bovine submaxillary mucin, which has higher sialic acid content had decreased interaction with S. aureus compared to pig gastric mucin. Heparin sulfate, a polyanionic polymer had even weaker interaction with S. aureus. By contrast, there was stronger interaction of S. aureus with methylcellulose, a neutral polysaccharide. Future studies using S. aureus strains with mutations in teichoic acid biosynthesis may be useful for further testing the effect of charge repulsion between mucins and the bacterial cell wall.
Advantages
Here, we report a quantitative, reproducible method for measuring the interaction of bacteria with mucus. This method could be adapted to studying a larger number of mutants or clinical isolates, and it could be adapted to study a variety of mucins or mucin analogs. Applying this method, we can identify S. aureus mutants with defective attachment to a standard surface and can quantitate the relative affinity of different S. aureus strains for mucin.
Limitations
The method we report does not capture the complexity of an in vivo or ex vivo airway epithelium, in which antimicrobial proteins, innate immune cells, and an abundance of macromolecules other than mucin may influence bacterial colonization. Furthermore, we did not replicate the cilia-directed flow and continuous secretion of mucin seen in airway infections. More specifically to mucin, pig gastric and bovine submaxillary mucins are standardized, commercially available mucin glycoproteins, but they have been denatured and do not retain their original viscoelastic properties. On the other hand, airway mucins are highly ordered (1, 2) and possess their native molecular properties. Our competition method for measuring interaction of mucin and bacteria is indirect and requires intact attachment of bacteria to the microtiter plate. However, because these mucins are easily washed off the plate surface, we found the indirect method was more feasible than direct interaction studies and did not require additional steps such as chemical modification of the mucins to conjugate them to a surface.
Conclusions
Our study shows that S. aureus attaches to strands of mucus. These strands are poorly cleared in CF. With a competition assay, we show that the affinity of S. aureus for mucin is relatively weak (10−2 mg/mL) and may be regulated by MgrA through repression of Ebh and SraP. This method will enable further study of bacterial factors involved the interaction with host mucins.
Author Contributions
Conceived and designed studies: AJF, ARH
Performed experiments: CPP, CRS, PDA, NA, ALC
Wrote first draft of paper: CPP, AJF
Edited paper: CPP, AJF, ARH
Supplemental Video 1. Confocal microscopy of excised pig trachea demonstrates reduced speed of S. aureus attached to mucus. Pig tracheas were treated with bumetanide while submerged in buffer lacking bicarbonate, slowing the emergence of mucus. Red fluorescent nanospheres were added to identify demonstrate mucus strands, which emerge from submucosal gland ducts. Green color indicates S. aureus that express green fluorescent protein. Unattached S. aureus are transported by ciliary flow. Many S. aureus are attached to the mucus strand. Their movement is slow compared to the unattached bacteria. Video is shown in real time.
Supplemental Video 2. Confocal microscopy of excised pig trachea shows an attachment event by S. aureus. Pig trachea was prepared similarly to the previous video, with red fluorescent nanospheres used to illuminate mucus strands. GFP expressing S. aureus were added. The video pauses at 10 s, when a new green particle of S. aureus appears immediately below the gland duct orifice. The video is slowed to 25% of normal speed until the S. aureus appears on the mucus strand (20 s). The S. aureus subsequently remains on the mucus strand.
Acknowledgements
This work was funded in part by NIH K08 HL136927 (AJF), CFF FISCHE16I0 (AJF), and NIH AI083211 (ARH). CPP was supported a fellowship for medical students NIH T35 HL007485 40. We thank Michael Welsh for his constructive feedback at several stages of this project.