Assessment of the Glycan-Binding Profile of Pseudomonas aeruginosa PAO1

ABSTRACT Pseudomonas aeruginosa is an opportunistic pathogen that can establish acute and chronic infections in individuals who lack fully functional innate immunity. In particular, phagocytosis by neutrophils and macrophages is a key mechanism that modulates host control and clearance of P. aeruginosa. Individuals with neutropenia or cystic fibrosis are highly susceptible to P. aeruginosa infection, thus underscoring the importance of the host innate immune response. Cell-to-cell contact between host innate immune cells and the pathogen, a first step in phagocytic uptake, is facilitated by simple and complex glycan structures present at the host cell surface. We have previously shown that endogenous polyanionic N-linked glycans localized to the cell surface of phagocytes mediate the binding and subsequent phagocytosis of P. aeruginosa cells. However, the suite of glycans that P. aeruginosa cells bind to on host phagocytic cells remains poorly characterized. Here, we demonstrate, with the use of exogenous N-linked glycans and a glycan array, that P. aeruginosa PAO1 cells preferentially attach to a subset of glycans, including a bias toward monosaccharide versus more complex glycan structures. Consistent with these findings, we were able to competitively inhibit bacterial adherence and uptake by the addition of exogenous N-linked mono- and disaccharide glycans. We discuss our findings in the context of previous reports of P. aeruginosa glycan binding. IMPORTANCE P. aeruginosa cells bind to a variety of glycans as part of their interaction with host cells, and a number of P. aeruginosa-encoded receptors and target ligands have been described that allow this microbe to bind to such glycans. Here, we extend this work by studying the glycans used by P. aeruginosa PAO1 cells to bind to phagocytic cells and by using a glycan array to characterize the suite of such molecules that can facilitate host cell binding by this microbe. This study provides an increased understanding of the glycans bound by P. aeruginosa and furthermore provides a useful data set for future studies of P. aeruginosa-glycan interactions.

IMPORTANCE P. aeruginosa cells bind to a variety of glycans as part of their interaction with host cells, and a number of P. aeruginosa-encoded receptors and target ligands have been described that allow this microbe to bind to such glycans. Here, we extend this work by studying the glycans used by P. aeruginosa PAO1 cells to bind to phagocytic cells and by using a glycan array to characterize the suite of such molecules that can facilitate host cell binding by this microbe. This study provides an increased understanding of the glycans bound by P. aeruginosa and furthermore provides a useful data set for future studies of P. aeruginosa-glycan interactions.
KEYWORDS Pseudomonas aeruginosa, glycan, array, phagocytosis, host-pathogen interactions phagocytes (9). These studies, which suggested that P. aeruginosa cells bind to clustered polyanions, guided our previous work, which revealed that endogenous negatively charged glycosaminoglycans (GAGs) can mediate P. aeruginosa adhesion. In turn, this observation led to the conclusion that N-linked glycans on phagocytes play a role in bacterial binding and phagocytosis (10). Therefore, a central goal of this study was to further elucidate the initial carbohydrate "handshake" between phagocytic cells and P. aeruginosa.
Glycans are found in all of nature and cover cell surfaces by decorating protein and lipid backbones (11,12). Glycans on host cells are often utilized by pathogenic bacteria for attachment and invasion (13)(14)(15)(16) and can even serve as a carbon/energy source (17). The binding of P. aeruginosa cells to glycans has been explored previously in several contexts. We showed that N-linked glycans and glycosaminoglycans on phagocytes can mediate the attachment and uptake of P. aeruginosa by macrophages (10). N-linked glycans have also been implicated as ligands for bacterial attachment to epithelial cells (18,19). P. aeruginosa cells can also bind glycan components of mucin (20)(21)(22)(23), complex glycans as part of glycolipids (24,25), and chitin (26,27).
In this work, we further explore the features of N-linked glycan structures that promote P. aeruginosa PAO1 binding, including mannose (Man), glucose (Glc), N-acetylglucosamine (GlcNAc), galactose (Gal), and fucose. To do so, we utilized single monosaccharide components of an N-linked glycan structure in competition assays. In parallel, we also performed a glycan array study to further evaluate the binding profile of P. aeruginosa to different classes of glycans. These studies have led to a better understanding of the glycans that can be bound by P. aeruginosa PAO1 and furthermore provides a useful data set for others investigating P. aeruginosa-glycan interactions.

RESULTS
Exogenous monosaccharides compete the phagocytosis of P. aeruginosa by THP-1 cells. We previously demonstrated that if N-linked glycan synthesis is inhibited on host phagocytic cells, the ability of P. aeruginosa cells to bind to the cell surface via these glycans is reduced (10). We hypothesized that the sugars that comprise N-linked glycan structures are necessary for the binding and uptake of P. aeruginosa. These N-linked glycan structures are primarily composed of mannose, N-acetylglucosamine, fucose, galactose, and sialic acid moieties, depending on the cell type; furthermore, N-linked glycans can be present in different quantities (28). There are three different classes of N-linked glycans (high mannose, complex, and hybrid), and each of these glycans share a core structure of Man 3 GlcNAc 2 bound to an asparagine (Asn) residue (29).
To assess the binding of glycan structures by P. aeruginosa cells, we performed competition assays in which we exogenously supplemented phagocytosis assays with components of N-linked glycans. In this set of experiments, we utilized P. aeruginosa PAO1, a motile strain of bacteria previously shown to interact with cell surface glycans (10,30), and performed gentamicin protection assays with human THP-1 monocytic cell lines, as reported (4,5,9,10), to assess the interactions of the microbe with host cells. The assays were performed for 45 min, which we showed previously allows for a reproducible measure of bacterial attachment and subsequent phagocytosis, with minimal change in the viability of the internalized bacteria (4,5,9,10).
We first assessed the phagocytosis of P. aeruginosa PAO1 cells by THP-1 cells in the presence of mannose (Fig. 1A) and 2a-mannobiose (Fig. 1B), a disaccharide of mannose linked by a 1-2 glycosidic bond (31). The addition of either of these compounds resulted in significantly reduced (;50%) uptake of P. aeruginosa. The addition of N-acetylglucosamine at a relatively high concentration (Fig. 1C) and fucose (Fig. 1D) caused a statistically significant but modest (;20%) reduction in P. aeruginosa phagocytosis. Finally, consistent with our previous observations (10), the internalization of P. aeruginosa was robustly and significantly diminished when the assay was supplemented with free galactose (Fig. 1E). Together, these data are consistent with the conclusion that P. aeruginosa PAO1 cells can bind to various mono-and disaccharide components that comprise high-mannose and complex glycans (Fig. 1F).
An assay to quantitatively profile interactions between P. aeruginosa and glycans. To identify additional candidate glycan ligands for P. aeruginosa, we utilized an otherwise wild-type (WT), green fluorescent protein (GFP)-expressing strain of P. aeruginosa PAO1 and performed a bacterial binding assay with the commercially available Glycan-100 array (RayBiotech). Each of the 100 glycan types was spotted in quadruplicate. An overnight culture of P. aeruginosa PAO1 carrying a GFP-expressing plasmid grown in lysogeny broth (LB) was subcultured and grown to mid-log phase in LB and then diluted 1:1,000 to ;1 Â 10 6 cells/mL in the manufacturer's minimal sample diluent buffer. After a 3-h incubation at room temperature with this bacterial suspension, each array was washed to remove nonbinding cells, and the fluorescence intensity values associated with the spotted glycans were quantified. These experiments were performed in triplicate; additional details of the experiment and its analysis are presented in Materials and Methods.
Pseudomonas aeruginosa binds to monosaccharides and an aminoglycoside. We further examined the binding results for each of the glycan classes spotted on the Glycan-100 array to identify the binding signatures. The glycans on the array can be divided into eight groups: (i) monosaccharides; (ii) aminoglycosides; (iii) disaccharides; (iv) blood groups, Lewis antigens, and fucosylated oligosaccharides; (v) globo-series glycolipids, milk oligosaccharides, and GAGs; (vi) natural oligosaccharides; (vii) gangliosides and sialylated oligosaccharides; and (viii) a-Gal and N-glycans. Notably, a majority of positive glycan signatures identified were to monosaccharides and disaccharides; 15 of the 28 significantly bound glycans fell into this category. This analysis also revealed that 14 out of the 28 glycans identified had at least one galactose moiety. Furthermore, 11 of the 28 glycans to which P. aeruginosa bound contained at least one GlcNAc moiety. We describe the binding of P. aeruginosa PAO1 cells to each specific glycan class below.
The significant attachment of P. aeruginosa cells to monosaccharides included (i) glucose, (ii) galactose, (iii) mannose, (iv) fucose, (v) rhamnose, (vi) GlcNAc, and (vii) GalNAc (Fig. 3A). With the exception of glucose and rhamnose, these glycans are all components of N-linked glycan structures and are capable of competing for bacterial binding sites on phagocytes (Fig. 1). Furthermore, P. aeruginosa cells significantly bound tobramycin (Fig. 3B), an antibiotic used for the treatment of this microbe.
In addition to N-linked glycans, P. aeruginosa PAO1 cells also bound to selected blood groups, Lewis antigens, and fucosylated oligosaccharides (

DISCUSSION
P. aeruginosa cells can bind to glycans in a variety of contexts. Previous studies have shown that glycan moieties contribute to P. aeruginosa binding of airway epithelium (18,19), and furthermore, bacterial surface adhesins can mediate binding to glycan ligands on host cells (32)(33)(34)(35) and on other bacterial species (13,15,36). In addition, recent studies have revealed that the inhibition of N-linked glycan synthesis on host innate immune cells by tunicamycin decreased bacterial binding and, consequently, phagocytosis of P. aeruginosa and interleukin 1b (IL-1b) elicitation from the host cells (10). However, studies to date have interrogated a limited number of glycans that do not encompass all of the possible complexity of glycan structures (10,11,36). In this work, we combined cellular assays based on knowledge gained from previous studies with a screen to identify additional candidate glycans bound by P. aeruginosa.
Since P. aeruginosa cells preferentially bind to N-linked glycans (10), as illustrated in Fig. 1F, our initial efforts were focused on identification of the components of these glycans that may confer binding. Competition studies for bacterial binding and uptake revealed that specific monosaccharide components of N-linked glycans, including mannose and galactose, are effective at competing bacterial association with THP-1 monocytic cells, while in comparison other monosaccharides (including fucose), at equimolar concentrations, Gal  are much less effective. We note an important caveat: while we added the bacteria and glycans immediately before the phagocytic assays, we cannot rule out that these additions also impacted the activation/physiology of the THP-1 monocytic cells.
To further identify the breadth of glycans bound by P. aeruginosa, we performed a quantitative binding assay with GFP-labeled P. aeruginosa cells using the Glycan-100 array. Interestingly, we found that the majority of bound glycans have one or more galactose moieties, which aligns with the competition studies presented here (Fig. 1). We note that galactose was found in the following glycans identified in the array assay: Gal-b-1,4-(Fuc-a-1,3)-Glc (glycan 25), blood A antigen (glycan 57), blood B antigen (glycan 59), GalNAc-b-1,3-Gal-b-1,4-Glc (glycan 20), Gal-b-1,4-GlcNAc-b-1,3-Gal-b-1,4-Glc (glycan 37), Glc-a-1,2-Gal-a-1,3-Glc (glycan 29), Neu5Ac-a-2,3-Gal-b-1,3-GlcNAc (glycan 11), Neu5Gc-a-2,6-Gal-b-1,4-Glc (glycan 24), Gal-a-1,3-Gal-b-1,3-GlcNAc (glycan 10), and SGP (glycan 100). Similarly, mannose, identified in the competition assays, was also bound by P. aeruginosa in the array study, and GlcNac was shown to mediate robust binding by P. aeruginosa, despite only a modest effect on phagocytosis in the competition assays (Fig. 1). Overall, from the eight specified classes Gal-a-1,3- (Fuc-a  a Glycan identities and mean fluorescence intensities (MFIs) of GFP-expressing P. aeruginosa PAO1 cells binding to 100 immobilized glycans. MFIs for each array were calculated as the average of the four data points from each glycan, spotted in quadruplicate on each array. A composite average MFI from the data across the 3 arrays (n 5 12 for each glycan moiety) is shown in the final column. Positive glycan partners for P. aeruginosa were identified by the MFI as being at least 100 RFUs over background and significantly greater than the negative control; 28 glycans met these criteria. Interaction of P. aeruginosa PAO1 with Glycans Microbiology Spectrum of glycans outlined above, monosaccharides, disaccharides, and an aminoglycoside were featured among the compounds bound by P. aeruginosa. Finally, we note that one limitation of this assay is that we cannot completely rule out some bacterial growth during the 3-h incubation from carryover of nutrients during the dilution of the culture 1:1,000 into the sample diluent buffer. Any such growth would simply enhance the relative binding signal, as growing bacteria that were not adhered to the array would be removed during the washing step. How do our findings fit within the context of previous studies? A previous study by Ramphal and colleagues showed that P. aeruginosa binds two different Gal-GlcNAc disaccharides (25), consistent with our findings here (Fig. 2) (compounds 9 and 53). Similarly, this microbe binds to the GalNAcb1-4Gal disaccharide (37), also consistent with our findings. Previous work also indicated that P. aeruginosa can bind N-acetylglucosamine and sialic acids (21), lipid-linked lactose and lactose-derivatives (24), and sialyl-Lewis X conjugates (22). We confirmed many of these findings here, supporting previous studies that showed the ability of P. aeruginosa to bind to specific moieties of glycolipids and mucins (20,25) and serving as a validation of our studies. P. aeruginosa has also been shown to bind to chitin (27), but we did not address binding to this sugar polymer.
One particularly intriguing observation was that P. aeruginosa binds to tobramycin (glycan 8). Tobramycin is a broad-spectrum antibiotic that is administered via intravenous or intramuscular injection and is also utilized heavily by oral inhalation for treatment of CF-associated chronic infections (38). Tobramycin inhibits protein synthesis by binding to the 16S rRNA of the bacterial 30S ribosome, which leads to mistranslation and subsequent cell membrane damage (39). This study suggests a novel mechanism of interaction between P. aeruginosa and tobramycin, perhaps explaining tobramycin's potency in killing this microbe (40). Additional studies will be needed to delineate how the observed binding may contribute to tobramycin's antibiotic activity.

MATERIALS AND METHODS
Bacteria. P. aeruginosa strain PAO1 carrying the GFP-expressing plasmid pSMC21 (41) was provided by D. Hogan and used here and in previous studies (4,5,42,43). Bacteria were cultured overnight in LB at 37°C and subsequently subcultured to achieve log-phase growth for 2 h in LB. Bacterial growth was determined by optical density at 600 nm.
Gentamicin protection assay. Phagocytosis of live bacteria was performed and quantitated as previously described (4,5,9,10). Briefly, overnight cultures of P. aeruginosa were washed, resuspended in serum-free Hanks balanced salt solution (HBSS; Corning), and added to the THP-1 cells at a multiplicity of infection (MOI) of 10. Where indicated, the phagocytic cells were treated with the specified glycan immediately before the competition studies, as previously described (10). The THP-1 cells were not activated prior to exposure to the bacteria or the glycans. After a 45-min coincubation of bacteria with the glycan or control, 100 mg/mL gentamicin was added to the assay for 20 min at 37°C. The cultures were washed and subsequently lysed in 500 mL 0.1% Triton X-100 solution in 1Â phosphate-buffered saline (PBS). Lysates were plated on LB plates and incubated overnight at 37°C to determine viable bacterial counts. The next day, recovered CFU on the LB plates were enumerated and represented as the percentage of the mean of WT bacteria phagocytosed or the fold increase in phagocytosis, as indicated in the figure legends, to quantitatively compare the relative phagocytosis levels.
Glycan-100 array analysis. The commercially available Glycan-100 array (RayBiotech) was used to assess carbohydrate binding preferences to 100 described glycan structures by P. aeruginosa. To assess glycan binding by P. aeruginosa, live GFP-expressing P. aeruginosa PAO1 cells were subjected to glycan array analysis (n 5 3). After growth overnight in LB medium, the GFP-expressing P. aeruginosa PAO1 cells were subcultured into fresh LB for ;2 h and then diluted 1:1,000 in the manufacturer's sample diluent buffer for a final bacterial cell count of ;1 Â 10 6 CFU/mL. This suspension was applied to the array and incubated at room temperature for 3 h as per the manufacturer's directions. The positive control was a fluorescent protein, and the negative control had nothing spotted on the array. Scanning and analysis of the slide were performed using the DNA Microarray Scanner (Agilent Technologies).
A list of the glycans on the Glycan-100 array is provided in Table 1. The glycan identities and mean fluorescence intensities (MFIs) of GFP-expressing P. aeruginosa PAO1 cells binding to the glycans were determined. The MFI for each glycan was calculated as the average of the four data points from each glycan (i.e., each glycan was spotted in quadruplicate on each array) from across the 3 arrays (thus, n 5 12 for each glycan moiety). This average MFI is shown in the final column of Table 1. Positive glycan-binding partners for P. aeruginosa were set using an MFI at .100 relative fluorescence units (RFUs) over background and significantly greater than the negative control; 28 glycans met these criteria.
Statistical analyses. The means 6 standard deviations (SD) derived from multiple independent experiments with technical replicates are shown for each graph. The sample sizes for each experiment are noted in the figure captions. As indicated in the captions, the unpaired Student's t test with Welch's correction or one-way analysis of variance (ANOVA) with Tukey's post hoc analysis were performed using Prism version 7.02 to determine the statistical significance of the data. Statistical significance is represented in the figures by asterisks.
Data availability. The glycan array is listed in Table 1, with the P. aeruginosa glycan-binding profile.