The bacterial strains used in the study and their origin are listed in Table 1 .

Preparation of full-length SP-D was performed as described (34). In brief, SP-D was purified by affinity chromatography from the culture supernatant of EXPICHO-S cells transfected with a plasmid encoding SP-D. The production of therapeutic-grade recombinant human MBL has been described previously (35). The recombinant proteins were subjected to reduced or native SDS-PAGE ( Supplementary Figure S2E ) and stained with Coomassie blue.

Recombinant human soluble Langerin ECD (residues 65-328 in Uniprot Q9UJ71) was expressed from the pET30a expression vector carrying kanamycin (KAN) resistance (GenScript, NJ). The vector was transformed into the Escherichia coli strain BL21 (DE3) (cat. no. EC0114, Thermo Fisher Scientific). Expression and purification were performed as described (36) with minor modifications. An overnight culture of transformed BL21 was diluted 40 times in Luria-Bertani media with 50 µg/mL kanamycin. The culture was grown at 37°C with agitation to an OD₅₅₀ ≈ 0.5, followed by 1 mM isopropyl-beta-D-thiogalactoside (IPTG) induction. After 3 h, the bacteria were harvested by centrifugation at 4,000 x g for 15 min at 4°C. The pellet was lysed by sonication (10 bursts of 30-sec durations, 45-sec rest) in 10 mM Tris-HCl, pH 7.5. Inclusion bodies were isolated by centrifugation at 10,000 x g for 15 min at 4°C and solubilized by sonication in 6 M guanidine-HCl and 0.01% beta-mercaptoethanol in 100 mM Tris-HCl, pH 7.5, followed by incubation on ice for 30 min. The mixture was centrifuged at 10,000 x g for 30 min at 4°C, and the supernatant was slowly diluted threefold in loading buffer (25 mM Tris-HCl, pH 7.8, 0.15 M NaCl, 25 mM CaCl₂) to allow protein refolding. The diluted mixture was dialyzed twice against 2 L loading buffer. The insoluble precipitate was removed by centrifugation at 10,000 x g for 30 min at 4°C.

The supernatant containing refolded recombinant human Langerin ECD was loaded onto Toyopearl HW 75f (TSK beads) as a pre-clearing step and incubated at 4°C for 1 h. The mixture was centrifuged for 5 min at 200 x g, and the supernatant was loaded onto mannose-coupled TSK beads, prepared as given in (37) and incubated for 2 h end over end. The beads were washed with loading buffer, and protein was eluted in EDTA-elution buffer (25 mM Tris-HCl, pH 7.8, 150 mM NaCl, 2 mM EDTA). The eluted fractions were analyzed by SDS-PAGE, and the protein concentration was determined through absorbance at 280 nm using a calculated molar extinction coefficient of 56,170 M^-1 cm^-1. Fractions containing the expected size of human Langerin ECD (~29 kDa) were pooled and further purified via size exclusion chromatography using a Superdex200 Increase 10/300 column (Cytiva) equilibrated in EDTA-elution buffer. Fractions in the peak signal were analyzed by SDS-PAGE, and fractions D7-E1 were pooled, as shown in Supplementary Figures S2A, B .

Neoglycoproteins were obtained from Dextra: fucose-BSA (Fuc-BSA) (cat. no. NGP1105), mannose-BSA (Man-BSA) (cat. no. NGP1108), N-acetyl glucosamine-BSA (GlcNAc-BSA) (cat. no. NGP1101), N-acetyl galactosamine-BSA (GalNAc-BSA) (cat. no. NGP1104), galactose-BSA (Gal-BSA) (cat. no. NGP1107) or BSA (cat. no. 9048-46-8, Sigma-Aldrich). These neoglycoproteins, soluble Langerin ECD, and MBL were dialyzed twice against phosphate-buffered saline (PBS), pH 7.4, and then once against PBS, pH 8.5. Biotin N-hydroxysuccinimide ester (cat. no. H1759, Sigma-Aldrich) was added (167 µg per mg protein) and reacted for 4 h at room temperature before termination of the reaction by three rounds of dialysis against Tris-buffered saline (TBS; 10 mM Tris, 140 mM NaCl, 0.09% (w/v) NaN₃, pH 7.4).

The Freestyle 293-F cell line (cat. no. R79007, Thermo Fisher Scientific), here referred to as HEK293, was cultured in 6-well plates to 70-90% confluency at transfection in Freestyle 293 Expression Medium (cat. no. 12338001, Thermo Fisher Scientific) with 10% fetal bovine serum (FBS) (cat. no. F7524, Sigma-Aldrich), 100 units/ml penicillin, and 100 µg/ml streptomycin (cat. no. 15140-122, Thermo Fisher Scientific). The HEK293 cell line was replenished with fresh expression medium with 10% FBS on the day of transfection. Expression constructs containing full-length open reading frames for Langerin (accession number NM_015717.5) and MGL (accession number NM_182906.4) were cloned into the mammalian expression vector pcDNA3.1/Zeo+ (GenScript), which confers resistance to Zeocin.

The Langerin- and MGL-encoding vectors and an empty vector (mock, i.e., the pcDNA3.1/Zeo+ vector alone) were introduced into the adherent HEK293 cell line (the cell line is adherent under the conditions used with FBS) by transfection. Briefly, DNA and Lipofectamine 3000 (cat. no. L3000001, Thermo Fisher Scientific) were mixed in two steps following the instruction manual. First, the lipofectamine 3000 reagent was diluted in OptiPRO serum-free medium (SFM) (cat. no. 12309050, Thermo Fisher Scientific). Second, 5 µg plasmid DNA was diluted in OptiPRO SFM, and then 10 µl P3000 reagent was added. The two mixtures were combined and incubated for 10 to 15 min, and DNA-lipid complexes were added to the cells by gently swirling the 6-well plate. The cells were replenished in a fresh culture medium after 24 h of incubation at 37°C and 8% CO₂. Forty-eight hours after transfection, the cells were supplemented with 500 µg/mL zeocin (cat. no. ant-zn-1, InvivoGen). The culture medium was changed every three days and analyzed for clonal outgrowth after 10-14 days of culture. Resistant clones were separated by limiting dilution into 96-well microtiter plates. Individual clones were screened for lectin expression by flow cytometry (see Confirmation of Langerin or MGL expression), and high expression clones were expanded for cryopreservation in freeze medium (293f Expression Medium with 20% FBS, 10% DMSO).

The HEK293-Langerin, HEK293-MGL, and HEK293-empty vector cells were thawed by resuspension in 293f Expression Medium with 10% FBS. Cells were sedimented by centrifugation at 200 x g and replenished in culture medium. Cells were incubated in standard T75 culture flasks (cat. no. 83.3911.002, Sarstedt) at 37°C and 8% CO₂. The next day, the culture medium was discarded and replaced with culture medium containing 500 µg/ml zeocin. The cells were split upon 80-90% confluency. Cells were harvested by incubating with PBS with 5 mM EDTA and immediately washed in HBS+ (20 mM HEPES, 150 mM NaCl with 5 mM CaCl₂, 5 mM MgCl₂, 0.5% BSA, at pH 7.4) by centrifugation at 200 x g for 5 min. Cell viability and numbers were assessed in the presence of trypan blue (cat. no T8154, Sigma Aldrich).

The expression of Langerin and MGL was confirmed by antibody staining. HEK293-Langerin, HEK293-MGL, or HEK293-empty vector cell lines (200,000 cells) were incubated with either 4 µg/mL anti-human CD207-APC (clone 10E2, BioLegend) or 6 µg/mL anti-human CD301-PE (clone H037G3, BioLegend) in HBS+. Subsequently, the cells were washed twice before flow cytometry on a NovoCyte 3000 flow cytometer equipped with three lasers (405 nm, 488 nm, and 640 nm) and 13 fluorescence detectors (Agilent, CA). All cells were gated using forward and side scatter height, whereas single cells were gated using forward scatter area and forward scatter height; see the gating strategy in Supplementary Figure S3A . Data were analyzed using FlowJO 10 (FlowJo, LLC).

The ability of the HEK293-Langerin, HEK293-MGL, or HEK293-empty vector cell lines to bind neoglycoproteins was examined using neoglycoproteins having ~20 monosaccharides coupled per BSA molecule (according to the manufacturer). The cell lines (200,000 cells) were incubated with 10 µg/mL biotin-conjugated neoglycoproteins (Fuc-BSA, Man-BSA, GlcNAc-BSA, GalNAc-BSA, Gal-BSA or BSA) in HBS+ for 30 min at room temperature. The cells were collected by centrifugation at 200 x g and washed twice before incubation with 0.5 µg/mL streptavidin-BV241 (cat. no. 563259, BD Bioscience) in HBS+ for 30 min at room temperature. The cells were subsequently collected by centrifugation and washed twice before flow cytometry was performed, as described above.

Green fluorescence protein (GFP) expressing S. aureus Reynolds (CP5), Reynolds (CP8), or Reynolds (CP-) were grown overnight at 37°C on Columbia agar with 2% (w/v) NaCl (CSA) or in suspension in Todd Hewitt broth (THB) at 37°C with shaking. The bacteria were harvested by centrifugation in the THB cultures or single colonies picked from the CSA plates. To inhibit non-specific binding, the bacteria were incubated in 1% heat-inactivated goat serum in HBS+ for 45 min at 4°C. Subsequently, the bacteria were incubated for 45 min at 4°C with 1 µg/mL rabbit anti-CP5 (38), 1 µg/mL rabbit anti-CP8 (38), or 10 µg/mL rabbit IgG (control) (cat. no. 7406404, Lampire biological laboratories) in HBS+, 1% goat serum and washed twice with HBS+, 0.2% goat serum. The bacteria were subsequently incubated with 2.5 µg/mL BV421 polyclonal goat anti-rabbit IgG (cat. no. 565014, BD Bioscience) for 45 min at 4°C and washed twice in HBS+, 0.2% goat serum before flow cytometry analysis with the Novocyte 3000 flow cytometer. The GFP-positive events were gated as bacteria; see the gating strategy in Supplementary Figure S3B . The data were analyzed using FlowJO 10.

We utilized the Nebraska transposon mutant library (32), comprising derivatives of S. aureus JE2, a community-associated methicillin-resistant USA300 LAC strain cured of three plasmids. The library consists of approximately 2,000 mutants, each with one nonessential gene disrupted by insertion of the mariner transposon bursa aurealis harboring an erythromycin (Em) resistance gene. The library includes transposon-mutants of tarS (NE0942) and tarM (NE0611), referred to as JE2ΔtarS and JE2ΔtarM. To generate a JE2ΔtarMS mutant, allelic exchange was performed as described in Bose et al. (39) to exchange the Em resistance cassette to Kan in JE2ΔtarS. To create the JE2ΔtarMS mutant, the ΔtarM mutation was transduced to JE2ΔtarS using φ80α with selection for Em resistance. Mutants were confirmed by their antibiotic resistance and by PCR using the primers listed in Table 2 . Bacterial growth rates and hemolysis on sheep blood agar plates were identical for the wild type (WT) and mutant strains.

GFP-expressing S. aureus Reynolds (CP5), Reynolds (CP8), or Reynolds (CP-) were cultivated overnight at 37°C on CSA or in THB. The bacterial suspensions were adjusted to 0.5 McFarland (McF) in HBS+. The HEK293-Langerin or -the empty vector cells (200,000 cells) were incubated with the GFP-expressing S. aureus strains for 30 min at room temperature. The cells were washed twice in HBS, 5 mM MgCl₂, 5 mM CaCl₂, and 0.5% BSA, and binding was measured using the NovoCyte 3000 flow cytometer. HEK293 cells were gated as described above, and the percentage of GFP-positive HEK293 cells was determined using FlowJO 10.

To examine the effect of the S. aureus CP on the binding of MBL or langerin ECD, GFP-expressing Reynolds (CP5), Reynolds (CP8), and Reynolds (CP-) were cultivated on CSA and adjusted to 0.5 McF in TBS, 5 mM MgCl₂, 5 mM CaCl₂, 0.5% BSA (TBS+) or TBS, 10 mM EDTA, 0.5% BSA. The bacteria were incubated with 5 µg/mL biotinylated langerin ECD or 1 µg/mL biotinylated MBL for 30 min, followed by a wash of the bacteria with either TBS+ or TBS, 10 mM EDTA, 0.5% BSA before incubation with streptavidin-BV421 in TBS+. The treated bacteria were subsequently washed twice. The BV421 signal was measured by flow cytometry using the NovoCyte 3000 flow cytometer and analyzed using FlowJO 10, where GFP-positive events were gated as bacteria.

The effect of the S. aureus CP on the interaction with SP-D was investigated by assessing bacterial aggregation induced by SP-D [protocol modified from (16)]. The degree of aggregation was measured by measuring light transmission through a bacterial suspension before and after the addition of SP-D. GFP-expressing S. aureus Reynolds (CP5), Reynolds (CP8), or Reynolds (CP-) were cultivated on CSA and adjusted to an OD₆₅₀ of 0.4 in TBS + 0.09% NaN₃. SP-D was prepared in either TBS, 5 mM MgCl₂, and 5 mM CaCl₂ or TBS and 10 mM EDTA buffers. Samples prepared in spectrophotometer cuvettes included 400 µL of the bacterial suspension and 100 µL SP-D (to a final concentration of 1 µg/mL) or a control with no SP-D. OD₆₅₀ was measured at the start of incubation and again after 16 h of incubation at 4°C. A decrease in optical density indicates aggregation of bacteria, which will sink in the cuvette and thus disappear from the optical path. The percent aggregation was calculated as ( 1 − O D 650   16   h O D 650   0   h ) · 100 .

To examine the influence of wall teichoic acids on lectin binding to S. aureus, strains JE2 WT, ΔtarS, ΔtarM, ΔtarMS, and ΔtagO were cultivated in THB at 37°C with agitation and stained with FITC. Overnight cultures were subcultured in fresh THB and cultivated to an OD600 ~0.6. The bacteria were washed by centrifugation and resuspended in 100 mM NaHCO₃ buffer, pH 9, containing 3.33 µg/mL fluorescein isothiocyanate (FITC) (cat no. F1906, Invitrogen). After incubation for 30 min at 37°C in the dark, the bacteria were washed twice in TBS. The mean labeling intensity of the bacteria is shown in Supplementary Figure S3C as determined by flow cytometry. The bacteria were resuspended in TBS+ to an OD₆₀₀ of 0.4, equivalent to ~10⁸ CFU/ml. HEK293-Langerin, -MGL or -empty vector cells (200,000 cells) were stained with a 1:2000 dilution of LIVE DEAD™ Fixable Near-IR Dead Cell Stain Kit (cat. no L10119, Invitrogen) for 30 min, washed, and incubated for 30 min at 37°C with the FITC-labelled bacteria at bacteria to cell ratios ranging from 1 to 9. Cells were subsequently washed in TBS+ and fixed in 1% formaldehyde in PBS for 15 min. S. aureus binding was measured using a Cytek Northern Lights flow cytometer (Cytek Biosciences, CA), a full spectrum flow cytometer equipped with three lasers (405 nm, 488 nm, and 640 nm). Unmixing of the spectral flow data was performed using a single stain of LIVE DEAD™ Fixable Near-IR Dead Cell Stain Kit of heat-killed cells, FITC-labelled bacteria, and BV421-FC beads (cat. No. 661627, BD Bioscience) using the SpectroFlo software (Cytek Biosciences). Unmixed data were analyzed using FlowJO 10.

Similarly, we examined the influence of the wall teichoic acid structure on the binding of soluble langerin ECD and MBL to fluorescent bacteria. S. aureus JE2 WT, ΔtarS, ΔtarM, ΔtarMS, and ΔtagO were cultivated and labeled with FITC as described above. Bacteria with an OD₆₀₀ of 0.4 in TBS+ or TBS, 10 mM EDTA, and 0.5% BSA were incubated with decreasing biotin-langerin ECD or biotin-MBL concentrations for 30 min at 37°C. Subsequently, the bacteria were washed twice and incubated with 0.5 µg/mL streptavidin-BV421 (BD Bioscience). The bacteria were washed and fixed in 1% formaldehyde in PBS for 15 min. The BV421 signal was measured using a Cytek Northern Lights flow cytometer. Unmixing of the spectral flow data was performed using FITC-labelled bacteria and BV421-FC beads, as described above.

GFP-expressing S. aureus Reynolds (CP-) was cultivated overnight on CSA at 37°C. Single colonies were picked and resuspended to 1 McF in HBS+. The bacteria were incubated with increasing amounts of MBL, SP-D, or human serum albumin (HSA) (cat. no. 109697, CSL Behring) for 20 min at room temperature, followed by incubation with either HEK293-Langerin or HEK293-empty vector cell lines for 30 min at room temperature. The cells were washed twice in HBS+. The signal from the GFP-labelled bacteria was detected using the NovoCyte 3000 flow cytometer, as described above.

For statistical analysis, normal distribution was assessed using qq-plots. A two-way analysis of variance (ANOVA) followed by Dunnett’s multiple comparison tests was performed in all cases where more than two groups were compared. Except in the SP-D aggregation assay, two-way ANOVA followed by Tukey’s multiple comparison test was used. Unpaired two-sided t-tests analyzed data to compare two groups. Before a t-test, an F-test was used to determine whether the variance was equal between groups. If there was equal variance, the data were analyzed by unpaired two-sided t-test; if not, Welch correction was utilized. Error bars are depicted as mean ± standard deviation (sd) in all bar graphs. Statistical analyses were performed using GraphPad Prism version 10. A p-value below 0.05 was considered significant.

A panel of different soluble and membrane-bound CTLs was generated and characterized ( Figure 1A ). For the membrane-bound CTLs, the HEK293 cell line was employed to generate cell lines constitutively expressing either langerin or MGL; a cell line transfected with an empty vector served as a negative control. The expression of langerin or MGL by the HEK293-cell line was confirmed by flow cytometry using specific antibodies ( Figure 1B, C ). The functional activity and specificity of the two membrane-bound lectins were characterized using BSA conjugated with different monosaccharides, i.e., neoglycoproteins. HEK293-Langerin cells exhibited significantly (P<0.0001) higher binding to GlcNAc-BSA, Man-BSA, and Fuc-BSA than to BSA. For the MGL-expressing HEK293-cells, we observed a highly significant (P<0.0001) specificity for GalNAc-BSA ( Figure 1D ). The specificities we find are consistent with the carbohydrate specificities reported in the literature (9, 10).

The langerin- and MGL-expressing HEK293 cells were used to screen for possible recognition of a range of different strains of S. aureus and Streptococcus pneumoniae. A collection of laboratory S. aureus strains (T1-T13 and Wood) was initially examined. HEK293-Langerin, compared to the empty vector control, showed binding of all strains ( Supplementary Figure S4 ). However, we observed a marked variation ranging from less than 20 percent to almost 80 percent of S. aureus-positive HEK293-Langerin cells. The MGL-expressing cells, on the other hand, did not bind any of the S. aureus strains examined. A collection of 89 different S. pneumoniae serotypes (i.e., representing 89 different capsular structures) and two nonencapsulated strains (ATCC12213 and C-mutant) was also screened for binding to HEK293-Langerin and HEK293-MGL. No binding of any of the S. pneumoniae strains was observed for either the HEK293-Langerin (despite good binding to the positive control S. aureus Wood) or the HEK-293-MGL cells ( Supplementary Figure S5 ). These results demonstrate that Langerin, but not MGL, can recognize S. aureus; neither Langerin nor MGL recognized any of the S. pneumoniae serotypes tested.

Others have reported the interaction between langerin and S. aureus (22). However, the majority of S. aureus strains produce CP, which is highly dependent on the bacterial culture conditions (40). CP is a staphylococcal virulence factor that is antiphagocytic and has been reported to mask surface antigens, evading recognition by PPRs of the innate immune system (26–29). Thus, we wanted to examine the influence of growth conditions and production of CP on the recognition of S. aureus by langerin.

S. aureus CP serotypes 5 and 8 represent at least 80% of clinical isolates (41). S. aureus Reynolds is a well-characterized strain for assessing the biological properties of CP; it produces CP5. To study the effect of CP on langerin binding, two isogenic mutants, a Reynolds strain producing CP8 and a strain with no CP production (CP-), were additionally employed (31). The degree of CP production was examined by the reaction of specific CP5 and CP8 antibodies with S. aureus strains cultivated either on CSA or in THB. On CSA, both Reynolds (CP5) and Reynolds (CP8) produced abundant CP5 and CP8, respectively, while Reynolds (CP-) did not produce CP ( Figure 2A ). When cultivated in THB, none of the three serotypes produced detectable amounts of CP ( Figure 2B ). We thus confirmed the observation that little CP is associated with staphylococcal cells in broth cultures (29, 42).

To study whether this difference in CP production influenced the interaction between langerin and S. aureus, GFP expressing Reynolds (CP5), Reynolds (CP8), and Reynolds (CP-) were cultivated on CSA or in THB. The bacteria were incubated with HEK293-Langerin, and the percentage of S. aureus-positive HEK293 cells was measured. When cultivated on CSA, both Reynolds (CP5) and Reynolds (CP8) showed a significant decrease in the interaction with HEK293-Langerin compared to Reynolds (CP-), from 58% S. aureus positive HEK393-Langerin cells to 11% for Reynolds (CP5) and 10% for Reynolds (CP8) ( Figure 2C ). Conversely, all three serotypes exhibited equal interaction with HEK293-Langerin when cultivated in THB. The results suggest that the CP effectively masks the surface of S. aureus cells for recognition by this innate immune system recognition molecule.

Hereafter, we assessed the influence of CP on the interaction of SP-D and MBL with S. aureus. Since the role of SP-D in the defense against pathogens is bacterial aggregation, we investigated the ability of SP-D to aggregate S. aureus. Bacterial suspensions of either Reynolds (CP5), Reynolds (CP8), or Reynolds (CP-) cultivated on CSA were incubated with SP-D in the presence of either calcium or EDTA, since the binding of SP-D is calcium dependent. The degree of aggregation was determined by measuring the light absorbance of the sample before and after 16 hours of incubation. As shown in Figure 3A , SP-D aggregated Reynolds (CP-) in a calcium-dependent manner (73% aggregation vs. 35% aggregation in the presence or absence of calcium, respectively). The encapsulated strains Reynolds (CP5) and Reynolds (CP8) failed to aggregate in the presence of SP-D.

The influence of CP on the interaction of MBL with S. aureus was investigated using flow cytometry. Reynolds (CP5), Reynolds (CP8), and Reynolds (CP-) cultivated on CSA were incubated with biotinylated MBL in the presence of either calcium or EDTA. MBL binding was detected by flow cytometry using streptavidin-BV421 to detect bound biotinylated MBL. As seen for bacterial aggregation by SP-D, MBL bound Reynolds (CP-) in a calcium-dependent manner, whereas MBL binding to both Reynolds (CP5) and Reynolds (CP8) was significantly decreased ( Figure 3B ). Although both SP-D and MBL interact with S. aureus in a calcium-dependent manner, bacteria that produce CP avoid aggregation by SP-D and recognition by MBL.

Finally, we examined the binding of a soluble form of langerin to S. aureus. Recombinant soluble langerin extracellular domain (langerin ECD) was expressed in E. coli and purified using mannose-coupled beads. This was followed by fractionation by size-exclusion chromatography (SEC). Through this procedure, we obtained a high-purity preparation of langerin ECD that eluted from the SEC column as a homogenous peak at ~90 kDa ( Supplementary Figures S2A, B ). This protein langerin ECD formed an expected trimer of ~90 kDa, as detected by mass photometry ( Supplementary Figure S2C ). Furthermore, the functionality of langerin ECD was examined, and as shown in Supplementary Figure S2D , langerin ECD bound mannose-BSA but not GalNAc-BSA or BSA. The langerin ECD was incubated in the presence of calcium or EDTA with Reynolds (CP5), Reynolds (CP8), or Reynolds (CP-) cultivated on CSA. We observed that langerin ECD binding was impaired by the presence of CP, i.e., a significant decrease in binding was observed for the encapsulated strains compared to Reynolds (CP-) ( Figure 3C ). Moreover, the addition of EDTA completely abolished langerin ECD binding ( Figure 3C ). The binding of langerin was blocked by the capsule, both when langerin was tested as a membrane-bound ( Figure 2C ) or as a soluble molecule ( Figure 3C ).

WTA has been shown to be a target on S. aureus for langerin and MBL (22, 23). WTA is an important constituent of the cell wall of Gram-positive bacteria. S. aureus WTA is a glycopolymer of ribitol phosphate repeating units modified with D-alanine and anchored to peptidoglycan ( Figure 4A ). Three glycosyltransferases, TarM, TarS, and TarP, are responsible for α-1,4-GlcNAc, β-1,4-GlcNAc, and β-1,3-GlcNAc modifications, respectively ( Figure 4A ). TarS is the most common glycosyltransferase in S. aureus. In some cases, TarS is co-expressed with TarM and, in rare cases, with TarP (25). It has been shown that langerin senses the β-GlcNAc but not the α-GlcNAc modification on WTA (17, 22), whereas MBL recognizes both modifications (23).

We utilized the S. aureus Nebraska transposon library (32) constructed in strain JE2, a capsule-negative isolate (43) in our experiments. JE2 NE0942 carries an insertion in tarS, and JE2 NE0611 carries an insertion in tarM, herein called JE2ΔtarS and JE2ΔtarM. To further examine the dependence of GlcNAc for langerin and MBL binding to S. aureus, a double mutant JE2ΔtarMS lacking both the TarS and TarM glycosyltransferases, was constructed. This strain lacks GlcNAc modifications of WTA since JE2 only expresses tarS and tarM. The JE2ΔtagO mutant lacks WTA since TagO is the enzyme responsible for initiating WTA biosynthesis (44).

The binding of MBL and langerin ECD to S. aureus JE2 WT and its WTA mutants in the presence of either calcium or EDTA was analyzed by flow cytometry. MBL binding was significantly increased for the JE2ΔtarM mutant compared to the WT strain, whereas there was an apparent loss of binding to the JE2ΔtarMS and JE2ΔtagO mutants ( Figure 4B ). This finding was confirmed when a range of different MBL concentrations in the binding reaction was tested ( Supplementary Figure S6A ). In contrast, langerin ECD showed only modest differences in binding to the WT strain JE2 and its WTA mutants ( Figure 4C ). Similarly, few differences were observed when different concentrations of langerin ECD were tested in the binding reaction ( Supplementary Figure S6B ). The JE2 ΔtagO mutant bound the langerin ECD in the presence or absence of calcium. These results are not consistent with previously published results (17, 22) since we observed only an indication of WTA-dependent langerin binding and suggest that langerin binds to an additional S. aureus ligand.

Previous reports measured the interaction between soluble recombinant langerin and S. aureus (17, 22). To measure the interactions between langerin-expressing cells and the different WTA mutants, HEK293-Langerin, HEK293-MGL, or HEK293-empty vector cells were incubated with FITC-labeled JE2 or its WTA mutants. Langerin binding to the HEK293 cells was expressed as the percentage of HEK293 cells that stained positive for S. aureus as detected by flow cytometry. Whereas embrane-bound langerin interacted with JE2 WT, JE2 ΔtarS, JE2 ΔtarM, JE2 ΔtarMS, and JE2 ΔtagO in a concentration-dependent manner ( Figure 4D ). No significant difference between the WT and the different mutants was observed. Consistent with our Langerin ECD experiments, our studies with HEK293-Langerin cells indicate an additional ligand for Langerin on S. aureus besides the β-1,4-GlcNAc modifications on WTA, as suggested by others (17, 22). HEK293-MGL did not exhibit binding to any of the S. aureus strains ( Figure 4D ).

Langerin, MBL, and SP-D all have comparable carbohydrate specificities, and our data indicate that they each interact with S. aureus. We addressed whether these similar CTLs compete for binding to S. aureus. GFP-expressing Reynolds (CP-) bacteria were incubated with HEK293-Langerin or HEK293-empty control with the addition of increasing concentrations of either MBL, SP-D, or human serum albumin (HSA). The percentage of S. aureus-positive cells was determined by flow cytometry. As shown in Figure 5 , both MBL and SP-D had inhibitory effects on the interaction between langerin and S. aureus, whereas HSA did not influence langerin binding. The data were fitted using four-parameter logistic regression to determine an IC50 for the two soluble CTLs. The IC50 for MBL was 0.407 nM, 95% CI [0.257, 0.557] and R² = 0.9210, whereas the IC50 of SP-D was 4.86 nM, 95% IC [4.164, 5.556] and R² = 0.9732. Thus, MBL was almost ten times better in the competition assay than SP-D. This might be due to a stronger interaction between MBL and S. aureus than SP-D and S. aureus. In conclusion, these data show that different lectins with similar affinities can compete for the same ligands in a complex natural setting.