Sialic Acid-Siglec-E Interactions During Pseudomonas aeruginosa Infection of Macrophages Interferes With Phagosome Maturation by Altering Intracellular Calcium Concentrations

Pseudomonas aeruginosa (PA) is commonly associated with nosocomial and chronic infections of lungs. We have earlier demonstrated that an acidic sugar, sialic acid, is present in PA which is recognized and bound by sialic acid binding immunoglobulin type lectins (siglecs) expressed on neutrophils. Here, we have tried to gain a detailed insight into the immunosuppressive role of sialic acid-siglec interactions in macrophage-mediated clearance of sialylated PA (PA+Sia). We have demonstrated that PA+Sia shows enhanced binding (~1.5-fold) to macrophages due to additional interactions between sialic acids and siglec-E and exhibited more phagocytosis. However, internalization of PA+Sia is associated with a reduction in respiratory burst and increase in anti-inflammatory cytokines secretion which is reversed upon desialylation of the bacteria. Phagocytosis of PA+Sia is also associated with reduced intracellular calcium ion concentrations and altered calcium-dependent signaling which negatively affects phagosome maturation. Consequently, although more PA+Sia was localized in early phagosomes (Rab5 compartment), only fewer bacteria reach into the late phagosomal compartment (Rab7). Possibly, this leads to reduced phagosome lysosome fusion where reduced numbers of PA+Sia are trafficked into lysosomes, compared to PA−Sia. Thus, internalized PA+Sia remain viable and replicates intracellularly in macrophages. We have also demonstrated that such siglec-E-sialic acid interaction recruited SHP-1/SHP-2 phosphatases which modulate MAPK and NF-κB signaling pathways. Disrupting sialic acid-siglec-E interaction by silencing siglec-E in macrophages results in improved bactericidal response against PA+Sia characterized by robust respiratory burst, enhanced intracellular calcium levels and nuclear translocation of p65 component of NF-κB complex leading to increased pro-inflammatory cytokine secretion. Taken together, we have identified that sialic acid-siglec-E interactions is another pathway utilized by PA in order to suppress macrophage antimicrobial responses and inhibit phagosome maturation, thereby persisting as an intracellular pathogen in macrophages.


INTRODUCTION
Sialic acid is a nine-carbon acidic sugar usually found at the terminal positions of carbohydrate chains which decorate many cell surface and secretory proteins of higher vertebrates. However, certain pathogenic bacteria such as Campylobacter jejuni, group B Streptococci and Neisseria synthesize sialic acids and utilize them to interact with sialic acid-binding immunoglobulin-like lectins (siglecs) on host cells leading to immune suppression for successful establishment of infection (1)(2)(3)(4). Other bacteria like Haemophilus influenzae are known to acquire sialic acids from their environment (5). We have previously demonstrated the presence of sialic acids on Pseudomonas aeruginosa (PA), a ubiquitous Gram-negative bacterium (6). These sialylated PA (PA +Sia ) interact with siglecs on neutrophils and reduce complement deposition (7). Such binding also impairs NETs formation, cytokine secretion, ROS generation as well as other biological functions thereby aiding their survival (8).
In general, macrophages respond to invading bacteria by secreting various cytokines, chemokines and attempt to eliminate such bacteria through phagocytosis followed by generation of reactive oxygen and nitrogen species (9, 10). Such phagocytosed bacteria are also trafficked into the endocytic pathway where acidification of phagosomal lumen, addition of degradative enzymes and fusion with the lysosome result in killing of most kinds of bacteria.
PA commonly infects immunocompromised and burn patients (11,12). They are usually found in lungs of cystic fibrosis patients as chronic infections (13). Alveolar macrophages play a critical role in managing such infections (14). PA produces secondary metabolites and several virulence factors which modify host cell responses during infection. In this context, it is worthwhile to investigate if sialic acid in PA +Sia can modulate macrophage immune responses against it by exploiting siglecs present on macrophages.
Here, we demonstrate that PA +Sia shows enhanced binding to macrophages through sialic acid-siglec-E interactions. Involvement of such sialic acid-siglec-E results in suppression of macrophage respiratory burst and reduced pro-inflammatory cytokines secretion. Additionally, this interaction resulted in the reduction of intracellular calcium ion concentrations during phagocytosis, modulating calcium-based cellular signaling thereby preventing the process of phagosome-lysosome fusion. Moreover, these sialic acid-siglec interactions recruited SHP-1/SHP-2 phosphatases at the immunoreceptor tyrosine-based inhibitory motifs (ITIM) of siglec-E. These phosphatases then interfere with MAPK, ERK, JNK-SAPK, and NF-κB pathways in PA +Sia -infected macrophages. Many of these observations were rescued by interrupting sialic acid-siglec-E interaction through silencing siglec-E. Taken together, we have established sialic acid as one of the important molecules utilized by PA to escape host innate immune response leading to its intracellular persistence in macrophages.

Acetyl Acetone Based Fluorimetric Estimation
As before, equal numbers of bacteria (5 × 10 9 ) were suspended in PBS and lysed by sonication in ice. Protein estimation was done and lysates (500 µg) were used to detect content of sialic acid by the method described by Shukla and Schauer (16). Mild oxidation of lysates by treatment with periodate solution (2.5 mM) at 4 • C in 15 min dark was followed by treatment with sodium arsenite (2% in 0.5 M HCl) and acetyl acetone (750 ml glacial acetic acid, 3.75 g ammonium acetate and 500 µl of acetyl acetone in 250 mL water) at 60 • C for 10 min then addition of 2.5 mL of water. Sialic acids present in lysates form a fluorogenic product after incubation and this product was quantified using a fluorimeter with excitation at 410 nm and collecting emission at a wavelength of 510 nm.

FITC-Labeling of Bacteria
Overnight cultured PA −Sia or PA +Sia was labeled with FITC as stated previously (7). Bacteria were pelleted, washed with PBS and its OD 600 value was determined. Equal numbers of PA −Sia or PA +Sia (1 × 10 9 ) were then incubated with 0.1% FITC in a minimal volume of bicarbonate buffer (pH 8.0) for 45 min at 37 • C. Unbound FITC was removed by repeated washing with PBS and extent of labeling with FITC was verified by flow cytometry.
Additionally, FITC-labeled bacteria were incubated for 0, 2, or 4 h in PBS at 37 • C followed by fixation with 2% paraformaldehyde to preserve FITC signal. Bacteria were analyzed by flow cytometry to detect stability of FITC tag.
Adhesion and Internalization of FITC-PA -Sia /PA +Sia With Macrophages by Confocal Microscopy J774A.1 cells (5 × 10 4 ) were seeded in glass coverslips and incubated overnight in IMDM medium at 37 • C with 5% CO 2 . FITC-PA was added to the coverslips at a ratio of 10:1 and incubated at 4 • C for 30 min. Unbound bacteria were removed by washing with PBS. These infected macrophages were finally fixed using 2% paraformaldehyde and mounted onto slides using DAPI containing mounting medium to visualize the binding of PA with macrophages by confocal microscopy. All images were acquired by Andor Spinning Disc Confocal microscope (Belfast, U.K.) with a 60 × /1.42 NA oil immersion objective.
In order to check the internalization of bacteria by macrophages, the above experiment was performed by incubating for 30 min at 37 • C followed by fixation and staining. Macrophages were also briefly treated with trypan blue solution (0.02%) to quench fluorescence coming from extracellularly adherent bacteria (17). Images were analyzed and bacterial fluorescence associated with individual cells was quantified by Fiji software (18). We have also calculated phagocytic index, which is the mean number of bacteria adhered or internalized per phagocytosing macrophage.

Phagosome-Lysosome Fusion in Infected Macrophages by Confocal Microscopy
Macrophages were similarly incubated with FITC-PA −Sia or PA +Sia at 10 multiplicity of infection (MOI) and incubated for 15 min at 4 • C to synchronize bacterial binding. Cells were then incubated at 37 • C for 30 min followed by removal of unbound bacteria by PBS wash. Infected macrophages were further incubated for 60 min. Lysotracker Red DND-99 (100 nM, L7528, Molecular Probes, Invitrogen) was then added and incubation continued for another 30 min at 37 • C to label lysosomes. Cells were fixed, mounted on slides, and analyzed by confocal microscopy as mentioned before.

Detection of Bacterial Localization in Endocytic Pathway by Confocal Microscopy
As previously described, coverslip-adherent macrophages were incubated with 1:10 MOI of FITC-PA −Sia or PA +Sia for 15 min at 4 • C. Plates were transferred to 37 • C for 10 min to allow synchronous uptake of bacteria and incubated further for 20 min, 40 and 60 min. Cells were fixed, permeabilized using PBS containing 0.3% Triton-X. Cells infected for 20 min were then overnight incubated at 4 • C with primary antibodies against Rab5 (1:100, #12666, Cell Signaling Technology) in blocking buffer (PBS containing 10% goat serum and 0.3% Triton-X). Cells infected with PA for 40 min were incubated with anti-Rab7 antibodies (1:100, #12666, Cell Signaling Technology) and those infected for 60 min were incubated with anti-LAMP1 antibodies (1:100, #3243, Cell Signaling Technology). Coverslips were washed with PBS containing 0.3% Triton-X and then counterstained with Alexa Fluor-647 tagged secondary antibodies (1:2000, A-21245, Invitrogen) in blocking buffer for 2 h at 4 • C. Coverslips were once again washed, mount and sealed. Images were randomly acquired from multiple fields in Fluoview FV10i (Olympus Life science). In order to confirm co-localization, multiple z stacks of 1 µm distance were also acquired. Images were processed and co-localization analyzed using JaCOP plugin (19) from Fiji software by calculating both Pearson's correlation coefficient (R) and Mander's overlap coefficient. For publication, images were brightened and background subtracted. On an average 100 cells from different fields were analyzed to determine colocalization.

Intracellular Calcium Levels in Infected Macrophages by Confocal Microscopy
Macrophages were also stained with Fluo-3-AM (5.0 µM, Sigma-Aldrich) for 30 min at 37 • C to allow the dye to permeate. They were then infected with PA −Sia or PA +Sia at 1:10 ratio, incubated for 15 min at 37 • C before fixation and processing for microscopy. Images were acquired in Fluoview FV10i (Olympus Life science). A similar experimental set up was repeated using siglec-E silenced macrophages for infection with PA +Sia . Images were analyzed as discussed above.

FLOW CYTOMETRIC ANALYSIS OF PHAGOCYTOSIS
Binding of PA by Macrophages J774A.1 (1 × 10 6 ) cells were incubated with 10 MOI of FITC-PA −Sia or PA +Sia at 4 • C for 30 min to detect their binding. Additionally, PA +Sia (1 × 10 6 ) were desialylated as mentioned earlier and used for binding.

Internalization of PA by Macrophages
Macrophages were incubated with 10 MOI of FITC-PA −Sia /PA +Sia /sialidase-treated PA +Sia at 37 • C for 30 min. As before, anti-siglec antibodies were used to block siglecs present on murine/human macrophages. Incubation was followed by washing to remove unbound bacteria. In order to distinguish between surface bound and internalized FITC-PA, cells were briefly treated with Trypan blue (0.02%) to quench FITC fluorescence coming from bacteria which were adherent on the surface. Thus, fluorescence associated with infected macrophages represents phagocytosed PA and was measured by flow cytometry as before.

Detection of Reactive Oxygen Species (ROS)
Macrophages (1 × 10 6 ) were infected with 10 MOI of PA −Sia /PA +Sia /sialidase-treated PA +Sia and incubated for 30 min at 37 • C. Non-adherent bacteria were removed by washing and cells were incubated with a ROS sensitive dye-CM-H 2 DCFDA (10 µM, C6827, Molecular Probes, Invitrogen) for an additional 30 min at 37 • C. Cells were washed and analyzed by flow cytometry to estimate ROS generation.

Quantitation of Nitric Oxide (NO)
Reactive nitrogen species (RNS) represented by NO, is another component of respiratory burst generated by macrophages in response to foreign particles. Intracellular nitric oxide was estimated using DAF-FM diacetate dye. As before, macrophages (1 × 10 6 ) infected with PA −Sia /PA +Sia /sialidase-treated PA +Sia were incubated with DAF-FM (100 nM, D23844, Invitrogen) for 30 min at 37 • C. Cells were washed and analyzed by flow cytometry.

Intracellular Survival of PA by Gentamicin Protection Assay
Intracellular survival and persistence of PA were assessed by gentamicin assay. Macrophages before and after treatment with anti-siglec antibodies were incubated with PA −Sia /PA +Sia /sialidase-treated PA +Sia at 10 MOI for 30 min at 37 • C. Unbound bacteria were removed by washing. Infected macrophages were further incubated for 3 h with gentamicin (200 µg/mL, G1397, Sigma-Aldrich) in IMDM medium to kill extracellular or surface adherent bacteria. To check the viability of internalized PA, macrophages were lysed in 1% w/v saponin solution. Lysate was serially diluted and spread on tryptic soy agar plates followed by overnight incubation at 37 • C. Viable and persistent PA gave rise to colonies and colony forming units (cfu) were used to calculate the number of viable bacteria inside macrophages.

Intracellular Proliferation of PA by CFSE Staining
Overnight cultured PA −Sia and PA +Sia were pelleted, washed with PBS and enumerated by spectrophotometry. Equal numbers of bacteria were incubated with CFSE dye (50 µM, C34554, Invitrogen) at 37 • C for 45 min to ensure uniform staining. Bacteria were repeatedly washed with PBS and then incubated with J774A.1 macrophages at 10:1 ratio at 37 • C for 30 min. Unbound bacteria were removed by PBS wash. Remaining uninternalized bacteria were killed by incubation with 200 µg/mL gentamicin. Following 8 h after incubation, cells were visualized by confocal microscopy to detect fluorescence from internalized CFSE stained bacteria.

Flow Cytometry Based Detection of Bacterial Localization in the Endocytic Pathway
J774A.1 (1 × 10 6 ) were incubated with 1:10 MOI of FITC-PA −Sia or PA +Sia for 15 min at 4 • C. Plates were then transferred to 37 • C to synchronize bacteria uptake and incubated for 20, 40, and 60 min. These cells were scraped off into a phagosome isolation buffer (20 mM HEPES, 0.5 mM EGTA, 0.25 M sucrose, 0.1% gelatin, protease inhibitor cocktail from Calbiochem, Merck-Millipore) (21). Cells were disrupted by passaging (15-20 times) through a 23-gauge needle fitted syringe. The suspension was centrifuged at 200 g for 10 min to separate most of the nucleus and intact cells. Phagosomes were collected by centrifugation of the supernatant at 13,000 g for 5 min and incubated for 5 min with 0.2% saponin, 10% normal goat serum in RPMI. Different endosomal compartments were stained by incubating with primary antibodies against Rab5 (1:100, #12666, Cell Signaling Technology), Rab7 (1:100, #12666, Cell Signaling Technology), and LAMP1 (1:100, #3243, Cell Signaling Technology) in the same buffer for 2 h at 4 • C. Phagosomes were counterstained with Alexa Fluor-647 tagged secondary antibodies (A-21245, Invitrogen). The localization of PA in various endosomal compartments was detected by flow cytometry (22).
Additionally, macrophages were infected for 45, 90, and 120 min, labeled with Lysotracker red dye and processed similarly as described above. The colocalization of FITC-PA into Lysotracker red positive lysosomal compartment was detected by flow cytometry.

Intracellular Calcium Concentration Estimation by Fluorescence Assay
J774A.1 (1 × 10 4 ) in microwell plates were incubated with Fura-2-AM (10 µM, Sigma-Aldrich) in presence of 0.04% pluronic-F in Hank's balanced salt solution (HBSS, pH 7.4) containing calcium (CaCl 2 0.14 g/L) at 37 • C for 30 min. Additionally, few cells were treated with ionomycin (10 µM) to estimate Rmax (high calcium). Similarly, cells were also incubated in calcium-free HBSS and treated with EGTA (10 mM) to estimate Rmin (low calcium). Fura-2-AM stained cells were infected with PA −Sia or PA +Sia at MOI 10 at 37 • C and intracellular calcium ions were detected by microplate fluorescence readers detecting fluorescence at 510 nm upon excitation both at 340 and 380 nm. Intracellular calcium concentrations in infected cells were estimated using reference wells as per formula developed by Grynkyewicz (23). Fluorescence readings were taken in Synergy TM 2 Multi-Mode Microplate Reader (Biotek Instruments Inc.).

Evaluation of Phagosome-Lysosome Fusion in Macrophages Upon Increase in Extracellular or Intracellular Calcium
J774A.1 macrophages were incubated in calcium-free Hank's balanced salt solution (HBSS) buffer or with additional 2.5, 5, 10 mM Ca 2+ (as CaCl 2 ). These cells were infected with 10 MOI of PA −Sia /PA +Sia at 4 • C for 30 min for bacterial binding followed by transfer to 37 • C. Cells were incubated for 30 min to allow phagocytosis in presence of extracellular calcium and then gentamicin (200 µg/mL) was added. Cells were incubated till 3 h followed by quantification of viable internalized bacteria by serial dilution and plating.
Additionally, J774A.1 cells incubated in calcium-free Hank's balanced salt solution (HBSS) buffer were infected with PA +Sia at 10 MOI for 30 min at 4 • C. Plates were transferred to 37 • C followed by addition of 5 mM Ca 2+ (as CaCl 2 ), ionomycin (10 µM) and were incubated further for 90 min to allow phagocytosis. Macrophages were also stained with Lysotracker red (100 nM) and processed for phagosome purification and flow cytometry as mentioned earlier (22). Macrophages were also prepared for confocal microscopy to detect phagosome lysosome fusion as noted before.

Immunoblot and Immunoprecipitation Analysis
J774A.1 macrophages (1 × 10 6 ) were infected with PA −Sia or PA +Sia at 10 MOI at 37 • C for 15, 30, and 45 min, respectively. Infected cells were scraped into PBS, sonicated in presence of protease and phosphatase inhibitor cocktail (Calbiochem, Merck -Millipore) and protein was estimated by BCA assay (Pierce, Thermo Scientific). Equal amounts of protein were resolved in SDS-PAGE (10%) gel and transferred on to a PVDF membrane via wet transfer method (24). Membranes were blocked and incubated with antibodies (1:1,000) against different signaling proteins for overnight. Membranes were next washed using Tris-buffered Saline containing 0.1% Tween-20 (TBS-T) then incubated with appropriate HRP-conjugated secondary antibodies and visualized using chemiluminescence substrate (Pierce, Thermo Scientific) in a Chemidoc imaging system (BioRad Laboratories, CA, USA). The primary antibodies Additionally, cell lysates were overnight incubated with antibodies against phospho-tyrosine proteins, SHP-1 or SHP-2 phosphatases. Proteins were immunoprecipitated from the lysate using avidin-agarose beads (Invitrogen) or protein-Asepharose 4B beads (Sigma-Aldrich). Beads were washed with chilled PBS to remove unbound proteins. Protein complexes bound to the beads were solubilized in sample buffer, resolved by SDS-PAGE, transferred to membrane and probed by western blotting process. Membranes were stripped off using stripping buffer (100 mM β-mercaptoethanol, 2% SDS, 62.5 mM Tris-HCl pH 6.8 for 30 min at 50 • C) and reprobed using the same antibody used for pulldown to check for protein loading.
Furthermore, cells were suspended into cytosol isolation buffer (10 mM Tris-Cl, 10 mM NaCl, 1.5 mM MgCl 2 , 1 mM PMSF, 0.05% NP-40, pH 6.8), vortexed to allow partial lysis of membranes, centrifuged at 1,000 g for 5 min to precipitate nuclear fractions and cytosol was collected as supernatant. The pellet was washed with chilled PBS, incubated in nuclear isolation buffer (20 mM Tris-Cl, 137 mM NaCl, 1 mM CalCl 2 , 1 mM MgCl 2 , 1 mM PMSF, 1% NP-40, pH 8.0) for 30 min. Nuclear fraction was collected after centrifuging the suspension at 1,000 g for 5 min at 4 • C. Both cytosolic and nuclear fractions were similarly resolved via SDS-PAGE and probed by western blotting.
Densitometric analysis of bands was performed using ImageJ software and values normalized with respect to β-actin band intensity have been represented as bar diagrams.

Knockdown of Siglec-E by siRNA Transfection
J774A.1 cells were transiently transfected with short interfering RNA (siRNA) designed against murine siglec-E (sc-153462, Santacruz Biotechnology) using Lipofectamine and Plus reagent (Invitrogen) as per manufacturer's instructions (15). Silencing of siglec-E was verified by western blotting. Siglec-E silenced J774A.1 cells were similarly infected with PA +Sia and macrophage response to infection was analyzed.

Statistical Analysis
The data represented are mean values derived from at least three independent experiments. Statistical analysis was performed using the two-tailed Student's t-test for two groups of samples and one-way analysis of variance (ANOVA) for more than two groups followed by pair wise multiple comparison procedures using Tukey test with p < 0.05 deemed statistically significant. Error-bars represent mean ± standard error of the mean (SEM) from three independent experiments. Significant differences were set at ns p > 0.05, * p ≤ 0.05, * * p ≤ 0.01, * * * p ≤ 0.001, and * * * * p ≤ 0.0001 and analyzed by GraphPad Prism version 6.01.

Sialylated PA Shows Higher Binding With Murine and Human Macrophages Through Sialic Acid-Siglec Interactions
We routinely confirmed the presence of sialic acid in PA before each experiment by sialic acid specific SNA, MAA lectin binding, and acetyl acetone based fluorimetric estimation of sialic acids.
Lectin binding was always higher with the sialylated (PA +Sia ) than non-sialylated bacteria (PA −Sia ) indicating presence of sialic acids in PA +Sia . This was further corroborated by sialidase treatment which abrogates this binding with PA +Sia . The MFI values of lectin binding have been presented in Figure 1A.
Additionally, sialic acid content from PA −Sia /PA +Sia /PA +Siasialidase was measured by fluorimetric acetyl acetone method (16). The estimated sialic acid content was always higher in case of PA +Sia as shown in Figure 1A.
Bacteria (PA −Sia /PA +Sia ) were labeled using FITC and analyzed by flow cytometry to confirm extent of labeling. No significant difference in labeling was observed between PA −Sia and PA +Sia despite presence of additional surface sialic acids ( Figure 1B). FITC tag was also found to remain stable till 4 h post-labeling ( Figure 1C).
Phagocytosis of FITC-PA +Sia and FITC-PA −Sia bacteria by human monocytic (THP-1), as well as murine (J774A.1) macrophage cell lines, were investigated. Phagocytosis involves the recognition and binding of the bacteria to macrophage surface through interactions between bacterial ligands and cellular receptors. This is followed by actual physical internalization by the phagocytic cell (9).
Additionally, PA +Sia were treated with sialidase to remove sialic acids. Desialylation of PA +Sia significantly abrogates (MFI = 63.67 ± 1.45, p ≤ 0.01) the higher binding seen in interactions of untreated PA +Sia with J774A.1 (MFI = 90.67 ± 0.88; Figure 2A). Thus, sialic acids on bacteria can be considered as ligands responsible for the differential binding between PA +Sia and PA −Sia to macrophages.
Sialic acids are recognized and bound by siglecs on host cells. In order to identify particular siglecs responsible for the increased association of PA +Sia with macrophages, cells were pretreated with blocking antibodies against siglecs. Flow cytometry data revealed that blocking of siglecs on J774A.1 macrophages with anti-siglec-E antibodies results in significant reduction in PA +Sia binding (MFI = 63.00 ± 0.57, p ≤ 0.05) as compared to untreated macrophages (MFI = 90.67 ± 0.88; Figure 2A).
Binding of PA +Sia to THP-1 derived macrophages (MFI = 322.7 ± 5.20) was similarly reduced when siglecs were blocked by anti-siglec 9 antibodies (MFI = 148.7 ± 4.05, p ≤ 0.05; Figure 2C). Treatment of macrophages with anti-siglec-5 antibody also slightly reduces the binding of PA +Sia (MFI = 272.7 ± 3.18, p ≤ 0.05). Representative histograms for the binding assays with J774A.1 cells are also shown in Figure 2B and with THP-1 is shown in Figure 2D. Blocking of other siglecs, namely  Figure S1A). Few other macrophage receptors are also known to mediate non-opsonic phagocytosis of PA (11). Contribution of such receptors in binding of PA +Sia was also checked by using anti-CD11b (CR-3 alpha chain) and anti-CD16/32 (FcγRII and FcγRIII) antibodies. Although these macrophage receptors are also involved in binding, siglecs also play an important role in PA +Sia binding with macrophages (Supplementary Figure S1A).
In order to further verify that PA entry into macrophages is due to its uptake by macrophages, cells were treated with cytochalasin D, which inhibits actin polymerization. Cytochalasin D nearly abolished bacterial uptake in both J774A.1 (MFI = 16.50 ± 0.50, p ≤ 0.05) and THP-1 derived macrophages (MFI = 23.33 ± 2.40, p ≤ 0.01; Figures 2E,G) confirming that macrophages are actively internalizing bacteria.
Taken together, these results demonstrate that PA +Sia binding to macrophages is mediated through sialic acid-siglec interactions mainly through siglec-E and siglec-9 subsequently leading to higher phagocytosis.

Internalization of PA +Sia by Macrophages Results in Poor Bactericidal Responses and Altered Cytokine Response
J774A.1 were infected with PA −Sia or PA +Sia for 30 min at 37 • C followed by staining with CM-H 2 DCFDA, an indicator dye for cellular ROS (Figure 3A). PA +Sia infected macrophages were found to exhibit significantly lowered ROS levels (MFI = 314.5 ± 7.81, p ≤ 0.0001) compared to PA −Sia infected macrophages (MFI = 513.7 ± 8.24). Interestingly, removal of sialic acids from PA +Sia resulted in enhancement of ROS production in macrophages Similarly, intracellular levels of nitric oxide generated were measured by staining infected macrophages with DAF-FM ( Figure 3B). It was observed that macrophages infected with PA +Sia show lower levels of intracellular nitric oxide (MFI = 255.2 ± 2.85, p ≤ 0.01) compared to PA −Sia infected macrophages (MFI = 401.6 ± 10.10). Removal of sialic acids from PA +Sia also resulted in enhanced generation of nitric oxide (MFI = 304.4 ± 18.10, p ≤ 0.05).
Thus, it may be noted that reduced respiratory burst and anti-inflammatory cytokine secretion in PA +Sia -infected  macrophages, can be reversed by desialylation of PA +Sia , suggesting that sialic acids-siglec interactions may play a role in this altered immune response.

Phagocytosed PA +Sia Evades Phagosome-Lysosome Fusion and Remain Viable Inside Macrophages
So far, we have observed that higher numbers of PA +Sia are internalized by macrophages relative to PA −Sia . However, attenuation of respiratory burst and anti-inflammatory cytokine responses by macrophages will have a negative impact on bacterial killing. Thus, the clearance of phagocytosed PA +Sia depends on the endocytic pathway, specifically phagosome maturation and phagolysosome formation. Accordingly, we checked the final fate of PA +Sia phagosomes inside the macrophages (Figure 4A). FITC-PA −Sia or PA +Sia infected J774A.1 macrophages were stained with Lysotracker red. Confocal microscopy revealed that across multiple fields, more numbers of green dots (intracellular PA −Sia ) colocalized with the red lysosomes giving a composite yellow color in the merged fields in contrast to PA +Sia -infected macrophages. Quantification of colocalization revealed that 98.7% of PA −Sia containing phagosomes colocalize with the lysosomal compartment while only 31.7% of PA +Sia containing phagosomes were found to do so.
We used flow cytometry to further quantify phagosomelysosome fusion. Lysotracker red stained macrophages infected with FITC-PA −Sia or PA +Sia for 45, 90 and 120 min were gently lysed and released phagosomes were analyzed in a flow cytometer. Out of all phagosomes, lysosomes and phagolysosomes were identified by incorporated lysotracker stain. The presence of FITC-PA in lysosomes, signifying phagosome-lysosome fusion, was quantified by dual color (FL1-FL2 channel) positivity ( Figure 4B).
In infected J774A.1, we observed that PA +Sia colocalized with lysosomes was lower being only 8.800 ± 0.30% (p ≤ 0.01) of all analyzed events compared to PA −Sia (17.90 ± 0.60%) after 45 min of infection ( Figure 4C). A similar trend was observed after 90 min of infection. The percentage of dual positive phagosomes diminished at 120 min post infection. We confirmed our observations in THP-1 derived macrophages, which showed a similar kind of colocalization at 45, 90 and 120 min upon infection with FITC-PA −Sia /PA +Sia (Figure 4D). Thus, the flow cytometry data corroborates with our observations from confocal microscopy, that fusion of PA +Sia -containing phagosomes with lysosomes was reduced.
Earlier, we demonstrated a comparatively lowered oxidative burst and reduced pro-inflammatory cytokines secretion (Figure 3) in PA +Sia -infected macrophages compared to PA −Sia . Furthermore, phagosome-lysosome fusion, which eliminates phagocytosed bacteria, is also reduced in PA +Sia -infected macrophages (Figure 4). Thus, we checked if PA +Sia bacteria remain viable following internalization by macrophages through gentamicin protection assay (Figure 4E). Results indicate that approximately 3.4-fold higher number of PA +Sia survived inside macrophages compared to PA −Sia ; colony forming units being 8.33 ± 0.43 × 10 5 vs. 2.90 ± 0.55 × 10 5 for PA +Sia vs. PA −Sia respectively (p ≤ 0.05) at 3 h post-infection.
Removal of sialic acids from PA +Sia resulted in ∼ >2.6-fold reduction in the number of viable bacteria isolated from infected macrophages. These observations clearly indicate that bacterial (E) Viability of bacteria in macrophages following phagocytosis was assessed using gentamicin protection assay. Macrophages were incubated with PA −Sia /PA +Sia /sialidase-treated PA +Sia at 1:10 MOI for 30 min at 37 • C followed by 3 h incubation with gentamicin (200 µg/mL). The viability and persistence of internalized bacteria were checked by plating the lysed macrophage contents on nutrient agar plates and checking for growing bacterial colonies (colony forming units, cfu). Data represented as mean ± s.e.m. of three independent experiments. (F) Macrophages infected with CFSE stained PA −Sia /PA +Sia were visualized by confocal microscopy to detect presence of bacteria (green specks). Scale bar = 5 µm Data represented as mean ± s.e.m. of three independent experiments. In all experiments, significance is represented as ns p> 0.05, *p ≤ 0.05, and **p ≤ 0.01.
sialic acid provide internalized PA +Sia with additional advantage to survive inside macrophages. So far we have observed higher numbers of viable PA +Sia inside macrophages after 3 h post infection by gentamicin protection assay. Staining of bacteria with CFSE is one of the methods to check bacterial multiplication inside phagocytic cells (25)(26)(27). Therefore, to check if internalized PA +Sia can replicate inside macrophages, J774A.1 macrophages were infected with CFSE stained PA −Sia /PA +Sia .
Macrophages infected for 8 h were visualized by confocal microscopy to check for the presence of PA ( Figure 4F). While only a few PA −Sia bacteria were observed as bright green dots inside the cells, multiple dots of faintly glowing PA +Sia were visible. This suggests that CFSE stained PA +Sia undergoes division inside macrophages which resulted in dilution of CFSE dye.

Reduced Recruitment of Rab7 Results in Inhibition of Phagosome-Lysosome Fusion in PA +Sia -Infected Macrophages
Phagosome maturation consists of early, mid and late stage phagosomes. Each stage is characterized by sequential association and dissociation of specific effector Rab proteins (28). The final step always is the fusion of the phagosome with lysosomes subsequently resulting in degradation of its cargo. Since phagosome-lysosome fusion was reduced in PA +Sia -infected macrophages, we wanted to identify if any intermediate step in phagosome maturation was hindered by checking for changes in the acquisition of stage-specific Rab proteins.
Previous experiments have demonstrated that internalized PA is trafficked into the lysosomal compartment within 45 min of infection. Accordingly, J774A.1 infected with PA −Sia /PA +Sia were fixed at 20, 40, and 60 min after infection, permeabilized and stained for early endosomal (Rab5), late endosomal (Rab7) and lysosome specific (LAMP1) markers, respectively. Cells were counterstained with secondary antibody and visualized by confocal microscopy. Nearly 100 cells from different randomly chosen fields as well as z stacks based visualizations were analyzed for colocalization of internalized bacteria with the endosomal compartments.
At the initial phase of infection (20 min), 79.67 ± 3.53% of internalized PA −Sia and 80.82 ± 6.15% of internalized PA +Sia were located in Rab5-stained compartment ( Figure 5A). Although PA −Sia and PA +Sia were equally localized in the Rab5 labeled early phagosomes, the number of PA +Sia taken up by the macrophages remained higher.
Maturation of early phagosomes into late phagosomes is marked by the appearance of Rab7 protein which is necessary for the final step of phagosome fusion to lysosomes (29). Accordingly, 40 min post-infected macrophages were stained with anti-Rab7 antibody. It was observed that 74.67 ± 9.61% of phagocytosed PA −Sia colocalized within the Rab7 stained compartment after 40 min of incubation ( Figure 5B). Interestingly, reduced numbers of PA +Sia -containing phagosomes (54.70 ± 11.31%, p ≤ 0.001) were found to acquire Rab7. Additionally, we compared colocalization between macrophages that had internalized higher numbers of PA −Sia and PA +Sia . Even at comparable numbers, PA −Sia always showed higher colocalization with Rab7 (Supplementary Figure S1B) while PA +Sia evaded fusion with the Rab7 compartment.
Next we wanted to observe successful fusion of PAcontaining phagosomes with lysosomes. Accordingly, lysosomes were labeled with anti-LAMP1 antibody in 60 min postinfected macrophages. Higher percentage of PA −Sia -containing phagosomes (81.72 ± 5.33%) was found to have reached the lysosomal compartments ( Figure 5C). However, at the same time, only 53.46 ± 14.23% (p ≤ 0.05) of PA +Sia -containing phagosomes were LAMP1 positive.
We also checked for colocalization of bacteria with the LAMP1 labeled compartment, which represents lysosomes and phagolysosomes (Figure 5F). At 20 min, the difference between PA −Sia and PA +Sia phagosomes was low mostly because very few phagosomes had matured at this early time point. We Frontiers in Immunology | www.frontiersin.org 13 February 2020 | Volume 11 | Article 332 FIGURE 5 | Alexa Fluor-647 tagged secondary antibodies and mount, sealed. Images were randomly acquired from multiple fields by confocal microscopy as well as from multiple z stacks. Images were processed and co-localization calculated using Fiji software. Colocalization results are calculated from minimum 100 cells from different fields. White arrows denote areas of colocalization between FITC-PA and labeled endosomes. Scale bar = 10 µm (D-F) J774A.1 were similarly infected with FITC-PA −Sia /PA +Sia for 20, 40, and 60 min. Cells were then processed as before and phagosomes were collected. Phagosomes were stained for Rab5, Rab7 or LAMP1 and analyzed by flow cytometry. Localization of FITC-PA was analyzed by dual fluorescence positivity. Data represented as mean ± s.e.m. of three independent experiments. In all experiments, significance is represented as ns p> 0.05, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.
observed that PA −Sia phagosomes associated with LAMP1 at 40 min was higher (3.050 ± 0.05%) than PA +Sia (2.350 ± 0.15%, p ≤ 0.05). At 60 min post-infection, a similar trend was observed. Colocalization of PA −Sia to lysosomes was 3.550 ± 0.05% while in case of PA +Sia this value was lower (2.450 ± 0.05, p ≤ 0.01). Taken together, our observations indicate that 20 min after infection, significantly higher numbers of PA +Sia was localized in Rab5 positive early phagosomes than PA −Sia . However, at the same time point, more PA −Sia had colocalized with the Rab7 compartment compared to PA +Sia . At 40 min after infection, higher numbers of PA +Sia continue to remain associated with early phagosomes while comparatively more PA −Sia -phagosomes mature and become Rab7 positive. Simultaneously, the PA −Sia colocalization with LAMP1 was higher than PA +Sia , signifying successful fusion to lysosomes. This observation was also confirmed by flow cytometry in FITC-PA −Sia and PA +Sia infected THP-1 derived macrophages (Supplementary Figure S1C).
Thus, our observations indicate that reduced acquisition of Rab7 by phagosomes harboring PA +Sia may be responsible for the subsequent prevention of phagosome-lysosome fusion as demonstrated by the reduced acquisition of LAMP1.

PA +Sia Modulates the Intracellular Calcium Level of Macrophages and Alters Calcium Related Signaling Thereby Affecting Phagolysosome Fusion
Modulation of intracellular calcium ions during phagocytosis has been linked to evasion of phagosome-lysosome fusion in Mycobacterium tuberculosis infections (30). Thus, we next evaluated intracellular calcium levels in PA −Sia or PA +Siainfected macrophages (Figure 6A). Microscopy revealed that intracellular calcium levels, [Ca 2+ ] c were lower in Fluo-3-AM loaded PA +Sia infected macrophages (7.075 ± 0.34 a.u., p ≤ 0.001) compared to PA −Sia (12.50 ± 0.70 a.u.) at 15 min post infection.
Additionally, intracellular calcium concentrations in Fura-2-AM stained infected macrophages were also analyzed by fluorescence plate reader and concentration of calcium ions (nM) was calculated using the Grynkewicz equation (Figure 6B). Following infection, intracellular calcium concentrations remained lower in PA +Sia -infected macrophages as compared to PA −Sia .
Alteration of intracellular calcium levels results in modulating a cascade of signaling events in the cytosol. Calmodulin is an important cytosolic calcium-binding protein that is known to regulate phagosome-lysosome fusion in Mycobacterium (30,31). Accordingly, we analyzed the levels of a few calciumrelated proteins in infected macrophages by western blotting (Figure 6C). Statistically significant differences were observed in levels of calmodulin, calcineurin, and calmodulin-dependent kinase type II expression only after 45 min post infection in PA +Sia infected macrophages.
So far we have observed that PA +Sia infection of J774A.1 macrophages are associated with lowered intracellular calcium concentrations and calcium-related proteins. As mentioned before, lowered calcium concentrations are correlated with phagosome lysosome fusion inhibition in Mycobacterium (30,31). Normally, infection was carried out in IMDM medium containing 1.5 mM calcium concentration (as CaCl 2 ). Gentamicin protection assay was used to determine if addition of extracellular calcium (final concentration −2.5, 5, 10 mM) affects the viability of PA +Sia at 3 h post infection. It was observed that bacterial viability was not affected in Ca 2+ free condition or at lower concentrations of extracellular calcium. However, viability of PA −Sia and PA +Sia was inhibited onwards from 5 mM extracellular Ca 2+ concentration ( Figure 6D).
We next checked if loss of viability of PA +Sia -infected macrophages in 5 mM extracellular Ca 2+ was due to enhanced phagosome-lysosome fusion. Macrophages in calcium free HBSS were infected with PA +Sia at 4 • C for binding. Plates were transferred to 37 • C and 5mM extracellular Ca 2+ ionomycin (10 µM) was added. Phagocytosis was continued in macrophages incubated in calcium-free HBSS or containing Ca 2+ (5 mM) or both Ca 2+ (5 mM) and ionomycin for 90 min. Cells were finally labeled with lysotracker red, phagosomes were prepared and analyzed by flow cytometry. Surprisingly, compared to infected macrophages in absence of calcium, addition of only extracellular Ca 2+ did not greatly increase phagolysosome fusion events ( Figure 6E). However, introduction of ionomycin along with extracellular Ca 2+ (5 mM) resulted in increase in the percentage of phagosome-lysosome fusion events. Ionomycin is known to increase intracellular calcium levels. Representative scatter plots indicating colocalization between lysotracker stained lysosome or phagolysosomes with FITC-PA +Sia are shown (Figure 6F).
Confocal microscopy also confirmed these observations. Phagosome-lysosome fusion remained unchanged at high extracellular Ca 2+ concentrations ( Figure 6G). But, addition of ionomycin increased the intracellular calcium concentration and enhanced phagolysosome fusion in PA +Sia infected macrophages ( Figure 6G).
These observations indicate that loss of viability of bacteria in presence of extracellular calcium was not related to phagolysosome fusion. However, in presence of extracellular calcium, ionomycin-mediated rapid increase in intracellular (Continued) FIGURE 6 | (D) Macrophages in calcium free Hank's balanced salt solution (HBSS) or supplemented with extracellular Ca 2+ at 2.5, 5, 10 mM concentration were incubated with PA −Sia /PA +Sia at 4 • C for 30 min to allow bacterial binding. Cells were then transferred to 37 • C for 30 min to allow phagocytosis and then gentamicin (200 µg/mL) was added. Cells were incubated till 3 h followed by quantification of internalized viable bacteria by plating. (E,F) Macrophages were incubated with PA +Sia in presence of enhanced extracellular Ca 2+ (5 mM) or intracellular Ca 2+ (5 mM Ca 2+ , 10 µM ionomycin) for 90 min at 37 • C. Phagolysosome fusion was detected in phagosomes released from these macrophages by flow cytometry as previously described. The results from three independent experiments are averaged and shown as bar diagrams with statistical significance. Scatter plots to detect double-positive populations are also shown. (G) Previous experiment was repeated and infected macrophages were observed by confocal microscopy to detect phagosome-lysosome fusion in infected macrophages in presence of enhanced extracellular or intracellular calcium. Images are representative of three independent imaging that was performed. (H) Cytokine response by PA −Sia /PA +Sia infected macrophages in presence of extracellular calcium of different concentrations was checked by ELISA. Levels of IFN-γ (pro-inflammatory cytokine) and IL-10 (anti-inflammatory cytokine) were estimated. Significance represented by ns p > 0.05, *p ≤ 0.05, and **p ≤ 0.01.
calcium enhanced phagosome-lysosome fusion in PA +Sia infected macrophages. This further confirms that intracellular Ca 2+ plays a regulatory role in bacterial killing through phagolysosome formation in PA +Sia -infected macrophages. Inflammatory cytokine response by macrophages incubated with PA −Sia or PA +Sia in presence of increased extracellular calcium was also checked by ELISA based estimation of IFN-γ (pro-inflammatory) and IL-10 (anti-inflammatory). No significant change in cytokine secretion was observed with increase in calcium ( Figure 6H). Interestingly, the differences in cytokine response between PA −Sia /PA +Sia -infected macrophages were diminished with increasing Ca 2+ .

Engagement of Siglec-E by Bacterial Sialic Acids Results in Siglec-Based Inhibitory Signaling in Macrophages
Binding of sialic acids by siglecs result in phosphorylation of tyrosine residues present in the ITIM motifs of inhibitory siglec-E leading to the recruitment of SHP-1 and SHP-2 phosphatases. We have already demonstrated that PA +Sia binds to J774A.1 through siglec-E (Figure 2). Immunoprecipitation revealed 2.75-fold higher phosphorylation of tyrosine residues of siglec-E in PA +Sia -infected macrophages compared to PA −Sia , indicating such engagement of siglec-E leads to its activation ( Figure 7A). A representative Ponceau blot is provided to account for equal loading across lanes. Next, we analyzed the association between siglec-E and SHP1/2 phosphatases (Figures 7B,C). Higher levels of siglec-E were found to be associated with SHP-2 in macrophages infected with PA +Sia compared to PA −Sia at each time point (20,40, and 60 min). At 40 min, association between siglec-E and SHP-2 was significantly ∼1.5-fold times higher (Figure 7C). These observations suggest that siglec-mediated inhibitory signaling is active in PA +Sia -infected macrophages. A similar trend was also observed with SHP-1 ( Figure 7B). Association between siglec-E and SHP-1/2 was also significantly higher at 40 min post infection in PA +Sia infected macrophages than PA −Sia infection.

Genetic Knockdown of Siglec-E Reveals Its Role in Modulating Macrophage Response Against PA +Sia
The role of sialic acid-siglec-E interaction on macrophage responses against PA +Sia was further established by silencing expression of siglec-E in J774A.1. Reduction in siglec-E expression due to transient siRNA transfection was confirmed by western blotting (Figure 8A). Consequently, confocal microscopy revealed 3-fold less PA +Sia were internalized by siglec-E knockdown macrophages (Figure 8B). Phagocytic index of PA +Sia infected macrophages was 11.62 ± 0.19 while it decreased to 6.467 ± 0.72 upon siglec-E knockdown.
We have observed that phagosome maturation is altered in PA +Sia -infected macrophages (Figures 4, 5). We have also observed modulation of intracellular calcium levels upon PA +Sia infection, which is known to play a significant role in such phagolysosome fusion inhibition (Figure 6). By microscopy, we identified that siglec-E has a role in such modulation, where we found higher intracellular calcium concentrations in siglec-E silenced macrophages (9.267 ± 0.29 a.u., p ≤ 0.01) compared to mock-transfected macrophages (6.833 ± 0.33 a.u.; Figure 9A) upon PA +Sia infection. This observation was corroborated by flow cytometry, where we observed that PA +Sia -infected siglec-E silenced macrophages show higher calcium levels (MFI = 1,994 ± 17.50, p ≤ 0.05) as compared to infected mocktransfected macrophages (MFI = 1,583 ± 92.00; Figure 9B). Additionally, we also demonstrated that calcium monitoring proteins such as calmodulin and calmodulin-dependent kinase II were also elevated upon siglec-E knockdown in macrophages ( Figure 9C). We have also confirmed that siglec-E knockdown resulted in reduction in expression of SHP-2 and SHP-1 in these PA +Sia infected transfected cells ( Figure 9C).
Statistical analysis of densitometry values of each band from three independent western blots were performed in all cases to evaluate the significance of observed differences in band intensities.
All these observations support that siglec-E plays an important role in regulating antimicrobial responses, intracellular calcium levels and key cellular signaling molecules in PA +Sia -infected macrophages.

Infection With PA +Sia Results in Altered Cellular Signaling
So far we have demonstrated the important role of siglec-E in sialic acid-siglec-E interactions in PA +Sia infected macrophages (Figure 9). Now, the status of a few key signaling molecules, previously identified to be affected by sialic acidsiglec-E interactions were also checked in PA −Sia /PA +Sia infected macrophages at different time points post-infection (Figures 9C,D). We have observed that expression of PI3K p85 and phosphorylated forms of ERK1/2, JNK/SAPK are consistently lower in case of PA +Sia -infected macrophages than PA −Sia infection ( Figure 10A). Average of densitometric values of observed protein bands normalized with respect to β-actin bands are reported as bar diagrams and statistical significance is also reported.
We also observed that the p65 subunit of NF-κB pathway remains in the cytosol in PA +Sia infected macrophages, resulting in lower activation of NF-κB pathway. However, p65 translocates into the nucleus of macrophages post infection with PA −Sia (Figure 10B) indicating activation of NF-κB pathway.

DISCUSSION
The rapid emergence of antibiotic-resistant strains makes treatment of PA infections quite challenging, causing worldwide concerns (32,33). Furthermore, PA possesses an arsenal of enzymes, virulence factors and a type III secretion system that play significant roles in its pathogenesis (34). Macrophages and neutrophils are innate immune cells which quickly respond to microbial infections by phagocytosing invading bacteria. In order to survive, several pathogens have devised methods to evade detection or prevent such killing by innate immune cells. Presence of sialic acid on PA has earlier been demonstrated as an important determinant for their survival against neutrophils (7). Therefore, we wanted to explore if sialic acids of PA have any impact on its interaction with macrophages.
The main finding of our current investigation includes a demonstration of higher binding and internalization of PA +Sia through interactions with siglec-E present on macrophages. Such sialic acid-siglec-E interaction exhibited reduced respiratory burst and anti-inflammatory cytokine secretion. More importantly, siglec-E plays an important role in modulating calcium signaling which leads to impaired phagosome maturation thereby interrupting phagosomelysosome fusion. Such sialic acid-siglec-E interaction also modulates several signaling molecules in MAPK and NF-κB pathways. Taken together, PA +Sia utilizes host immune receptor (siglecs) strategically to impair macrophage immune response leading to their survival and persistence inside the macrophage.
In case of PA infections in lower airways, alveolar macrophages uptake these bacteria mostly by non-opsonic phagocytosis through complement receptor-3 and CD14 receptors for their elimination from the lung (12,35). We have observed similar binding of PA +Sia to macrophages through complement receptor-3 (Supplementary Figure S1). Most importantly, we have identified that PA additionally utilizes sialic acid for higher binding and increased internalization inside both human and murine macrophages. Desialylation of PA +Sia reduced this enhanced binding indicating a critical role of sialic acids (ligands) in such binding.
Macrophages eliminate phagocytosed bacteria by chemokine and cytokine secretion, phagosomal acidification, ROS generation and finally phagolysosome formation (9). Oxygen represented as mean ± s.e.m. (C) Expression of MAPK, ERK, JNK pathway molecules signaling molecules and calmodulin, calmodulin-dependent kinase type II was detected in lysates from PA +Sia -infected mock or siglec-E siRNA transfected macrophages by western blotting. (D) Similarly, infected macrophages were separated into cytosolic and nuclear fractions. Proteins were resolved in SDS-PAGE and probed for the expression of NF-κB signaling pathway to detect nuclear translocation of p65 subunit. Western blots (C,D) were representative of three independent experiments. Densitometric values reported are normalized with respect to β-actin band intensities and such values from atleast three independent experiments are used to calculate the mean band intensities which are presented as bar diagrams alongwith statistical significance of observed changes. Significance represented by *p ≤ 0.05 and **p ≤ 0.01. and nitrogen radical generation by macrophages effectively eliminates PA in cystic fibrosis patients as well (41). TH1 type cytokine responses lead to better clearance of PA in the lungs and higher survival rates of infected mice (42).
Here we report that antimicrobial responses such as ROS and RNS generation were suppressed in PA +Sia -infected macrophages. This corroborates our previous report that PA +Sia interactions with siglec-9 on neutrophils also result in a similar reduction in respiratory burst (7). Also, PA +Siainfected macrophages show an anti-inflammatory cytokine response, characterized by higher levels of IL-4, IL-12, and TGF-β secretion. Thus, macrophage response against PA +Sia is impaired. However, interruption of sialic acid-siglec interactions by desialylation of PA +Sia or blocking or silencing siglec-E on macrophages results in reversal of such innate immune responses.
Following internalization, bacteria bound in phagosomes may undergo fusion with lysosomes. A lysosome is a highly acidic organelle containing enzymes like lysozymes, proteases, lipases and cathepsins along with redox stressors responsible for breakdown of invading microbes, exogenous particles, apoptotic bodies or damaged cellular organelles (9, 43). We followed the trafficking of PA +Sia phagosomes through the different stages of phagosome maturation using both microscopy as well as flow cytometry. At initial time points, higher numbers of PA +Sia were localized in early endosomes (Rab5 positive), consistent with our observations that more number of PA +Sia were internalized. However, fewer bacteria were found in the late phagosomal compartment (Rab7 positive) at both 20 and 40 min post infection (Figures 5B,E). Since Rab7 mediates the final fusion with lysosomes (labeled by LAMP1), inhibition of Rab7 recruitment may have caused lowered phagosome maturation in PA +Sia -infected macrophages (Figure 4). Taken together, lowered respiratory burst, antiinflammatory cytokines and inhibition of phagosome-lysosome fusion lead to impaired killing of PA +Sia . This was confirmed by gentamicin protection assay where PA +Sia remained viable following 3 h after internalization inside the macrophages. Interestingly, we have observed that PA +Sia also undergo cellular division inside the macrophages through dilution of CFSE dye (Figure 4).
Current research has raised questions about the generalization that PA is only an extracellular pathogen as it has also been found to exist intracellularly (44)(45)(46). The toxin ExoS has been implicated in inhibiting phagosome-lysosome fusion in infected epithelial cells, allowing PA to survive intracellularly in membrane blebs (46)(47)(48). ExoS also interacts with Rab5 thereby preventing phagocytosis of bacteria (49). A study involving human and mouse primary macrophages has also reported that PA is capable of entering, surviving as well as multiplying intracellularly in macrophages, depending on expression of bacterial outer membrane protein-OprF (50). Other factors from PA such as mgtC and oprF allow PA to persist intracellularly in macrophages (45,51). There are several indications that PA can influence endocytic pathways for their intracellular existence. Here, we have observed that macrophages fail to completely eliminate PA −Sia or PA +Sia . However, we have demonstrated that PA +Sia utilizes bacterial surface sialic acids to interact with siglecs on macrophage surface as a strategy to enhance its internalization as well as intracellular survival.
The phenomenon of phagosome-lysosome fusion inhibition has been extensively studied in a few bacteria like Mycobacterium sp., Salmonella enteric serovar typhimurium, and Francisella tularensis (52). Decrease in intracellular calcium concentrations and altered calcium signaling results in inhibition of phagosomelysosome fusion in Mycobacterium infection (30). However, the role of intracellular calcium in modulating phagosomelysosome fusion is controversial (30,31,(53)(54)(55)(56)(57)(58). In an in vitro system, calcium has been demonstrated to be essential for fusion between late endosomes and lysosomes (59). Furthermore, enhanced intracellular calcium also induces phagosome-lysosome fusion and inhibits bacterial viability (30). In murine macrophages interacting with unopsonized yeast cells, phagosome maturation is reported to be regulated by intracellular calcium concentrations (58). Particularly, phagocytosis of PA also depends on the increase in intracellular calcium concentration mediated by TRPV2a calcium channel on macrophage plasma membrane (60). Here, we also demonstrate that a reduction in intracellular calcium concentration accompanied by lower expression of calcium-sensing molecules is correlated with the inhibition of phagosome maturation during PA +Sia infection (Figures 4,  6). Also, ionomycin-mediated increase in intracellular calcium actually enhanced phagosome-lysosome fusion rate in PA +Siainfected macrophages (Figure 6) indicating that intracellular calcium concentration modulates phago-lysosome fusion in PA +Sia -infected macrophages.
Increasing the extracellular Ca 2+ concentration also lead to reduction of viability of internalized PA −Sia /PA +Sia but the mechanism of bacterial clearance was not due to phagosomelysosome fusion. Interestingly, extracellular calcium abrogates the anti-inflammatory response observed in PA +Sia -infected macrophages. Together with the reduction in viability of internalized PA +Sia , this might suggest that extracellular calcium, at higher concentrations, may improve anti-bacterial response by macrophages.
Certain calcium-sensitive proteins like Calmodulin and calmodulin type kinase II play an important role in phagolysosome fusion in Mycobacterium infected macrophages. Calmodulin is known to be involved in phagosome maturation and phagolysosome formation both in cell-free systems and bacteria-infected cells (30,(61)(62)(63)(64).
Calmodulin also plays a role in antimicrobial responses by associating with inducible nitric oxide enzyme (iNOS) and enhancing nitric oxide production in macrophages (65). In the case of PA +Sia -infected macrophages, lowered calmodulin levels, and reduced nitric oxide production hints at this relation. In a calcium-dependent manner, calmodulin activates calcineurin, which interacts with a group of transcription factors like the calcineurin-dependent nuclear factor of activated T cells (NFAT). Calcineurin-NFAT signaling is also known to collaborate with NF-κB signaling and induces pro-inflammatory cytokine production (66). Here we report that calcineurin expression was lowered in PA +Sia -infected macrophages from 45 min post infection, which may impact such calcineurindependent signaling events leading to an anti-inflammatory cytokine response.
Thus, calcium signaling and related pathways are modulated due to sialic acid-siglec-E interactions. Reduced intracellular calcium levels as well as calcium-sensing proteins calmodulin and calmodulin kinase II were reversed by silencing siglec-E suggesting an important contribution by siglec-E in modulating calcium in PA +Sia -infected macrophages.
These results indicate that siglec-E can possibly function like siglec-G (CD22) on B cells which inhibits calcium flux from B-cell receptor (BCR) ligation through recruitment of SHP-1 phosphatases (67). The involvement of siglec-8 in mast cells similarly inhibits calcium flux generated due to FcεRI cross-linking (68). Rise in intracellular calcium ions in macrophages during PA infection may be due to TRPV channel recruitment (60) or PA −Sia binding with other receptors (69)(70)(71) which is consequently diminished in PA +Siainfected macrophages. We have demonstrated an increased phosphorylation of tyrosine residues of siglec-E and enhanced association of siglec-E and phosphatases SHP-1/SHP-2 in PA +Siainfected macrophages. In a similar fashion, siglec-E mediated SHP-1/2 recruitment in case of PA +Sia infection may be responsible for lowered intracellular calcium levels in contrast to PA −Sia . Taken together, we may propose that this sialic acidsiglec-E interaction modulate macrophage intracellular calcium levels thereby decreasing phagosome-lysosome fusion in PA +Siainfected macrophages.
A contrasting report by Perdicchio et al. indicated that sialylation does not affect the intracellular routing of antigens hence artificially sialylated ovalbumin reached lysosomes like unmodified ovalbumin (72). However, virulence factors of PA are known to modify host cellular responses, including modulating the endocytic pathways (51). Interactions between such a sialylated bacterial pathogen with macrophages are fundamentally different from that involving artificially sialylated antigen uptake by bone marrow derived dendritic cells (BM-DCs).
The impact of siglec-E-sialic acid interactions on cellular signaling was convincingly demonstrated by silencing siglec-E which reduced PA +Sia association with macrophages but restored the antimicrobial response via high ROS, RNS generation and pro-inflammatory cytokine secretion. Additionally, MAPK pathway and NF-κB translocation were also modulated by such silencing.
Taken together, we have demonstrated that PA +Sia are internalized more by both murine and human macrophages. Such sialic acid-siglec interactions influence anti-bacterial responses like respiratory burst, modulation of intracellular calcium flux and calcium related signaling molecules. All these FIGURE 11 | PA +Sia modulates macrophage responses to promote its survival inside macrophages through sialic acid-siglec-E involvement. The schematic diagram summarizes the altered macrophage immune response against PA +Sia due to involvement of siglec-E which leads to persistence of PA as an intracellular pathogen. events lead to altered cytokine response through MAPK, ERK and NF-κB signaling pathways and ultimately inhibition of phagosome-lysosome fusion. As a result, the phagocytosed PA +Sia are comparatively eliminated less and found to persist intracellularly in the macrophages establishing the critical role of sialic acids and siglec-E for bacterial persistence. Our findings regarding the role of sialic acid-siglec-E interactions in PA pathogenesis of macrophages is illustrated in a schematic diagram (Figure 11). Therefore, the potential payback of our work is a greater understanding of the roles of sialoglycans and the function of siglec-E in interactions between sialylated bacteria and immune cells as a way to better understand bacterial pathogenesis.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/Supplementary Material.