CD56-mediated activation of human natural killer cells is triggered by Aspergillus fumigatus galactosaminogalactan

Invasive aspergillosis causes significant morbidity and mortality in immunocompromised patients. Natural killer (NK) cells are pivotal for antifungal defense. Thus far, CD56 is the only known pathogen recognition receptor on NK cells triggering potent antifungal activity against Aspergillus fumigatus. However, the underlying cellular mechanisms and the fungal ligand of CD56 have remained unknown. Using purified cell wall components, biochemical treatments, and ger mutants with altered cell wall composition, we herein found that CD56 interacts with the A. fumigatus cell wall carbohydrate galactosaminogalactan (GAG). This interaction induced NK-cell activation, degranulation, and secretion of immune-enhancing chemokines and cytotoxic effectors. Supernatants from GAG-stimulated NK cells elicited antifungal activity and enhanced antifungal effector responses of polymorphonuclear cells. In conclusion, we identified A. fumigatus GAG as a ligand of CD56 on human primary NK cells, stimulating potent antifungal effector responses and activating other immune cells.


Introduction
Invasive pulmonary aspergillosis (IPA), most commonly caused by the opportunistic mold pathogen Aspergillus fumigatus, is a devastating infection in immunocompromised patients.Individuals that are highly susceptible to IPA often have neutropenia and/or dysfunctional Tcell responses [1].This includes patients suffering from hematological malignancies, undergoing allogeneic hematopoietic stem cell transplantation (HSCT) [2,3], with genetic predisposition [4], or suffering from severe respiratory illnesses like Influenza or Covid-19 [5].IPA poses a major clinical challenge due to limited diagnostic tools, insufficient efficacy of available antifungal therapies, and increasing antifungal drug resistance [3,6,7].Hence, IPA is associated with high mortality rates [2,8] and poor long-term survival [9].Thus, understanding host defense and host-pathogen interaction are pivotal for the development of diagnostic biomarkers, targeted anti-virulence agents, and novel immunotherapies.
Several studies have shown that natural killer (NK) cells are an integral part of protective antifungal immunity [10,11].For instance, HSCT recipients with delayed NK-cell reconstitution or low NK-cell counts are at higher risk of developing IPA, indicating that NK cells are indispensable for fungal clearance [12].NK cells constitute 5-15% of the blood lymphocyte repertoire and can either secrete regulatory cytokines to recruit and stimulate other immune cells or elicit cytotoxicity against target cells [13,14].Several mechanisms of NK-cell-mediated cytotoxicity against tumor cells and pathogen-infected cells have been identified.While elimination of target cells through death receptor-mediated apoptosis are critical for NK-cell activity against tumor cells [14][15][16], this mechanism does not appear to be the main driver of the cells' antifungal activity [17].Instead, different studies emphasized the importance of lytic granules, containing perforin, granzyme B and granulysin, as mediators of NK-cell cytotoxicity against fungal pathogens [18][19][20] Besides their direct killing capacity, NK cells are able to release a number of chemokines (CCL3, CCL4, CCL5) and cytokines (IFN-γ) by which they modulate and augment the activity and host response of a number of immune cells such as neutrophils and macrophages [18,21,22].

PLOS PATHOGENS
CD56 -Aspergillus fumigatus interaction is triggered by galactosaminogalactan Innate immune cells, including NK cells, are commonly activated by stimulation of pattern recognition receptors (PRR).These receptors detect pathogen-associated molecular patterns (PAMPs) of invasive pathogens.Several fungal-reactive PRRs of NK cells have been described.For instance, the natural cytotoxicity receptor (NCR) NKp30 was shown to recognize β-1,3-glucan on the surface of Candida albicans and Cryptococcus neoformans [23], while the activating human NK-cell receptor NKp46 interacts with the Candida glabrata adhesins Epa1, Epa6, and Epa7 [24].Furthermore, we previously identified the NK-cell receptor CD56 as a PRR recognizing A. fumigatus.Upon contact with A. fumigatus hyphae, CD56 relocalizes actin-dependently toward the NK-cell/hyphal interaction site, leading to prominent reduction of CD56-fluorescence intensity.Blocking of CD56 resulted in decreased A. fumigatus-induced NK-cell activation and chemokine secretion, suggesting a functional relevance of CD56 in hyphal recognition by NK cells [25].However, while PAMPs for the NK cell receptors NKp46 [24] and NKp30 [23] have been identified, the A. fumigatus ligand of CD56 has remained unknown.
Given our prior observation that CD56 solely interacts with A. fumigatus germ tubes and hyphae but not with conidia, we specifically focused on hyphal PAMPs to identify the A. fumigatus ligand of CD56.Using a combination of purified cell-wall components, biochemical treatments, and A. fumigatus mutants, we herein identified galactosaminogalactan (GAG), especially de-N-acetylated GAG, as the fungal ligand of CD56.The GAG/CD56 interaction triggered strong NK-cell activation, along with potent release of cytotoxic effectors and immune-enhancing chemokines.Furthermore, supernatants of GAG-pulsed NK cells inhibited fungal growth and enhanced the anti-Aspergillus activity of PMNs, suggesting GAG as a potential immunotherapeutic target in IPA.

CD56-mediated recognition of A. fumigatus depends on intact cell wall polysaccharides
In a first step, we sought to determine whether the binding partner of CD56 on A. fumigatus hyphae is a cell wall protein or polysaccharide.Therefore, we depleted proteins in the cell wall of A. fumigatus germ tubes by proteinase K treatment before co-culture with NK cells and flow cytometric quantification of CD56 expression after fungal stimulation.Relocalization of CD56 toward the immunological synapse with A. fumigatus is accompanied by a remarkable decrease in fluorescence intensity [25].NK cells confronted with either untreated A. fumigatus or with proteinase K-treated A. fumigatus displayed a significant reduction in CD56 fluorescence positivity.(Fig 1A), suggesting that cell-wall-bound proteins are not the ligands of CD56.
Next, we treated A. fumigatus germ tubes with periodate that oxidizes cell wall polysaccharides having vicinal diols and co-cultured them with NK cells.Compared to NK cells co-cultured with untreated A. fumigatus, flow cytometry revealed no decrease in NK-cellular CD56-fluorescence intensity when challenged with periodate-treated germ tubes (Fig 1B).This suggests impaired relocalization of CD56 on NK cells co-cultured with periodate-treated

PLOS PATHOGENS
CD56 -Aspergillus fumigatus interaction is triggered by galactosaminogalactan germ tubes.To reinforce these findings, bulk RNA sequencing was performed to identify genes that are differentially regulated in NK cells upon exposure to periodate-treated versus untreated A. fumigatus germ tubes.NK cells challenged with periodate-treated A. fumigatus showed 43 significantly differentially expressed genes compared to cells exposed to untreated germ tubes (Fig 1C).Out of these 43 genes, 38 genes (28 with a >0.5 log 2 fold change) had weaker expression in NK cells stimulated with periodate-treated fungus, including genes associated with NKcell activation (e.g., CD69) and chemokine secretion (e.g., CCL3 and CCL4) (Fig 1C and 1D).
Next, we confirmed that declined CD56 interaction of A. fumigatus after periodate treatment is dependent on the duration of periodate exposure (Fig 1E).Furthermore, periodate treatment time-dependently reduced the ability of A. fumigatus germ tubes to trigger secretion of macrophage inflammatory protein (MIP)-1α (CCL3), MIP-1β (CCL4), RANTES (CCL5) and perforin from NK cells compared to untreated germ tubes (Fig 1F).Altogether, these results suggest that CD56 on human NK cells interacts with a hyphal cell wall polysaccharide and not a protein.

Binding of CD56 to A. fumigatus cell wall galactosaminogalactan triggers NK-cell activation
To identify the specific cell wall polysaccharide that binds to CD56, we used an enzyme-linked immunosorbent assay (ELISA), in which different cell-wall polysaccharides extracted from the A. fumigatus hyphal cell wall were coated on microtiter plates and then incubated with recombinant CD56.Although CD56 interacts only with germ tubes and hyphae of A. fumigatus, we also tested binding to conidia-specific cell surface components, RodA protein (RodAp) and melanin pigment.Among the tested compounds, CD56 only bound with high affinity and dose-dependently to galactosaminogalactan (GAG), especially its urea-insoluble fraction (Figs 2A and S1A) [29].
Next, we performed a pull-down assay to assess CD56 binding to GAG compared to β-1,3-glucan, the major cell wall polysaccharide of A. fumigatus (Fig 2B).After interaction with GAG or β-1,3-glucan, CD56 could be extracted from the GAG pellet (lane A) but not from β-1,3-glucan pellet (lane E).Instead, CD56 was found entirely in the supernatant after co-incubation with β-1,3-glucan (lane D).Collectively, these findings suggest a direct and specific interaction between CD56 and GAG.
GAG is a heteropolysaccharide composed of α-1,4-linked monomers of galactose, N-acetylgalactosamine (GalNAc), and galactosamine (GalN), synthesized during germination of A. fumigatus [30,31].Our CD56 ELISA suggested that CD56 particularly binds to the urea-insoluble GAG fraction (PGG) enriched in GalNAc (Fig 2A).To confirm the specificity of the interaction between PGG and CD56, PGG was coated on ELISA plates and probed with soluble CD56 that has been pre-incubated with acid-hydrolyzed PGG.Indeed, this hydrolysate markedly reduced the binding affinity of recombinant CD56 to PGG (S1B Fig), corroborating that the interaction between PGG and CD56 is specific.D = donor, fc = fold change.(D) Volcano plot summarizing transcriptional changes in NK cells stimulated with periodatetreated A. fumigatus germ tubes compared to cells stimulated with untreated germ tubes.Genes with a log2 fold change > 0.5 and an adjusted p-value < 0.05 are highlighted.Selected immune-related genes have been labelled.(E) CD56 expression on naïve NK cells (Control) and NK cells stimulated for 3 h with A. fumigatus germ tubes, depending on the duration of periodate pretreatment of the germ tubes.Representative histograms and data for cells from 5 independent donors are shown.(F) Chemokine and perforin release by naïve NK cells (Control) and NK cells stimulated for 3 h with A. fumigatus germ tubes, depending on the duration of periodate pre-treatment of the germ tubes.Previous studies found significant differences in cell wall-bound GAG among Aspergillus species.Specifically, A. fumigatus harbors the highest amount of cell wall-associated GAG, whereas A. nidulans produces low quantities of GAG [31].Therefore, we compared the decrease in CD56 surface expression and NK-cell activation after exposure to A. fumigatus and A. nidulans at different multiplicities of infection (MOIs).NK cells co-cultured with A. fumigatus showed stronger reduction in CD56 fluorescence intensity at all tested MOIs than NK cells confronted with A. nidulans (Fig 2E).Likewise, A. fumigatus induced stronger surface expression of CD69 and CD107a as well as chemokine secretion than A. nidulans, especially at a low MOI (Fig 2E and 2F).These results indicate that the quantity of cell wall-associated GAG plays a role in CD56-mediated NK-cell activation.

CD56 binds to wild-type A. fumigatus hyphae but not to mutants lacking GAG
As purified GAG (i.e., PGG) from A. fumigatus triggered strong NK-cell activation and cytokine release, we next investigated the potential of the GAG-deficient mutant strains Δagd3 [32] and Δuge3 [33] to interact with CD56 on NK cells.Therefore, we either co-incubated wildtype (WT) A. fumigatus (strain Af293) and the two otherwise isogenic mutants Δuge3 and Δagd3 (lacking key enzymes for GAG biosynthesis) with soluble CD56 or bovine serum albumin (BSA) protein as control, followed by staining with fluorescent anti-CD56 antibody.Soluble CD56, strongly bound the hyphal surface of A. fumigatus WT, but not the hyphal surface of either GAG-deficient mutant (Fig 3A).Upon incubation of A. fumigatus WT or the mutants with BSA instead of soluble CD56, we found only dim autofluorescence of the hyphae, ruling out any unspecific binding of the fluorescent antibody to hyphal structure (insert in Fig 3A).In summary, these findings indicate specific binding of CD56 only to the cell wall of A. fumigatus WT hyphae that produce functional GAG.

A. fumigatus mutants lacking GAG fail to elicit NK-cell activation
To further strengthen the role of GAG in CD56-mediated hyphal recognition and its potential to induce NK-cell activation, we co-cultured NK cells with A. fumigatus WT (Af293) or the

GAG is indispensable for CD56 accumulation at the site of A. fumigatus/ NK-cell interaction
To further corroborate that CD56 on NK cells binds to A. fumigatus GAG, we incubated NK cells with A. fumigatus WT hyphae or hyphae of the GAG-deficient strain Δagd3 and performed confocal laser scanning microscopy (CLSM) to assess CD56 localization.Consistent with our previous study [25], NK cells co-cultured with the WT strain displayed a strong CD56 signal at the contact site, whereas NK cells exposed to the deacetylase-deficient Δagd3 mutant maintained a homogenously distributed CD56 fluorescence signal across the plasma membrane (Fig 3D) [32].This confirms that CD56 concentrates at the fungal interface in the presence of its ligand GAG.

Deacetylated residues of GAG are crucial for interaction with CD56 and NK-cell activation
Given the absence of CD56 binding to Δagd3 mutant hyphae, we further sought to test the importance of deacetylation on the CD56-GAG interplay.We first analyzed CD56 binding to either fully N-acetylated GAG (aPGG), fully de-N-acetylated GAG (dePGG), and native GAG (PGG), using ELISA.In contrast to native GAG, no binding of aPGG to CD56 was observed, whereas dePGG showed intermediate binding to CD56 (Fig 4A).Incubation of PGG, aPGG, and dePGG with only the secondary antibody or a combination of anti-CD56 and secondary antibody in the absence of recombinant CD56 yielded only a weak background signal, confirming that binding of CD56 to (de)PGG is specific (Fig 4A ).
Next, we assessed the capacity of aPGG, dePGG, and native PGG to activate human NK cells.Consistent with the ELISA results, we observed a significant decrease in CD56-fluorescence intensity on NK cells stimulated with native PGG but not with aPGG, while dePGG followed by staining with fluorescent anti-CD56 antibody.The shape of the hyphae is indicated by dotted lines The insets represent a control image (bottom; BSA stained with anti-CD56 antibody) of the respective strain.Scale: 10 μm.(B) CD56, CD69, and CD107a expression on naïve NK cells (Control) and NK cells stimulated for 6 h with WT Af293 or the GAG-deficient A. fumigatus mutants (Δuge3 and Δagd3) at different multiplicities of infection (MOIs).Representative histograms for one donor at MOI 4 are shown.(C) MOI-dependent induction of NK-cellular secretion of granzyme B, perforin, and chemokines after 6-h stimulation with WT Af293, Δuge3, and Δagd3.(B-C) N = 3 independent donors.Repeated measures (RM) one-way analysis of variance (ANOVA) with Dunnett's post-hoc test versus Control, i.e., unstimulated NK cells (asterisks).In addition, results for stimulation with the 3 strains at each MOI was compared using RM one-way ANOVA with Dunnett's post-hoc test versus WT (hash signs).To further elucidate the specificity of the interaction of CD56 with deacetylated galactosamine residues of GAG, we stimulated NK cells with GAG oligomers exclusively consisting of GalN or GalNAc residues.Also, considering the involvement of galactosamine in the interaction with CD56, we investigated the binding affinity of CD56 to chitosan, a polymer composed of de-N-acetylated glucosamine and N-acetyl-glucosamine units [34,35].Both, GAG and chitosan are cationic polymers due to their deacetylation [32,34,35].Interestingly, only NK cells pulsed with GalN oligosaccharides displayed a significant reduction in CD56 fluorescence intensity, along with upregulation of CD69 and CD107a surface expression (Fig 4D).In contrast, GalNAc oligomers were unable to interact with CD56 and induce NK-cell activation.Likewise, chitosan showed neither binding to CD56 nor modulation of CD69 or CD107a (Fig 4D).Consequently, and in contrast to GalN oligomers, GalNAc and chitosan did not stimulate the secretion of chemokines CCL3, CCL4, and CCL5 (Fig 4D).Together, these findings indicate that the presence of galactose residues is not required for the interaction of CD56 with GAG.Moreover, the biochemically similar, positively charged polymer chitosan did not trigger NK-cell activation, further corroborating the specificity of the interaction between CD56 and GAG.

The complete, GalN-rich GAG molecule is required for interaction of CD56 with A. fumigatus
To further validate that the complete, GalN-rich GAG molecule is required for CD56-mediated NK-cell activation, we preincubated germ tubes of A. fumigatus WT (Af293) or Δagd3 with the GAG-specific carbohydrate-active enzymes Sph3 (hydrolase degrading GalNAc homo-oligomers [36]), Ega3 (hydrolase degrading GalN homo-oligomers [37]), and/or Agd3 (deacetylating homo-oligomers [32,38]).As shown above, without enzyme pretreatment, only A. fumigatus WT resulted in a prominent reduction in NK-cellular CD56 fluorescence intensity (Fig 5A).Addition of either of the hydrolases or a combination of both to the NK-A.fumigatus co-culture alleviated A. fumigatus (WT)-induced CD56 relocalization, corroborating CD56 binding to the complete, GalN-rich GAG molecule.Expectedly, pretreatment of Δagd3 germ tubes with Agd3, allowing for external complementation of the mutation and GAG deacetylation, significantly decreased NK-cellular CD56 fluorescence intensity to a similar extent as non-pretreated A. fumigatus WT.As seen with A. fumigatus WT, this effect can be alleviated by addition of hydrolases.As previously observed, all conditions that led to decreased CD56 fluorescence intensity in turn induced enhanced CD69 expression (  Altogether, these data suggest that the intact, GalN-rich GAG molecule is required for CD56-mediated NK-cell activation and release of antifungal effector molecules.

Supernatants of PGG-stimulated NK cells inhibit fungal growth and stimulate polymorphonuclear neutrophils (PMNs)
Next, we sought to assess whether the effector responses induced by CD56/GAG interaction exhibit any anti-A.fumigatus activity.Therefore, we used an IncuCyte time-lapse fluorescence microscope and the NeuroTrack image processing algorithm [39] to monitor fungal growth and hyphal branching.Compared to A. fumigatus alone, supernatants of NK cells stimulated with PGG significantly inhibited hyphal growth and branching (Fig 6A).In contrast, supernatants of unstimulated NK cells from the same donors did not inhibit fungal growth and morphogenesis (Fig 6A).
In addition to their direct inhibitory effect on fungal growth, we evaluated the capacity of supernatants from PGG-activated NK cells, which contain increased levels of neutrophil-activating chemokines CCL3 and CCL4, to engage PMNs and enhance their anti-Aspergillus activity.Expectedly, PMNs partially inhibited fungal growth and hyphal branching of A. fumigatus in our IncuCyte NeuroTrack assay.Interestingly, supernatants from PGG-stimulated NK cells significantly enhanced suppression of hyphal growth and morphogenesis by PMNs (Fig 6B), even at a 2.5-fold lower concentration of supernatants from PGG-stimulated NK cells than in our previous experiment (Fig 6A ), a concentration that per se did not inhibit mycelial proliferation.Additionally, we analyzed A. fumigatus-induced ROS production and phagocytosis of A. fumigatus-GFP conidia.While supernatants of unstimulated NK cells did not impact neutrophilic ROS response to the germ tubes, PGG-pulsed NK-cell supernatants significantly enhanced Aspergillus-induced ROS release of PMNs (Fig 6C).Consistently, supernatants of PGG-pulsed NK cells had a stronger stimulatory effect on phagocytosis of A. fumigatus-GFP by PMNs than supernatants of unstimulated NK cells (Fig 6C).Altogether, supernatants of PGG-stimulated NK cells elicited both direct antifungal activity and an indirect immunostimulatory effect through engagement of PMNs

Discussion
While alveolar macrophages and PMNs form the first line of antifungal defense in the lung, NK cells are an integral part of second-line defense against A. fumigatus.Several studies showed an involvement of CD56 in NK-cell-mediated cytotoxicity [40][41][42].For instance, the deletion of CD56 resulted in failed polarization during immunological synapse formation [43].We have previously shown that CD56 serves as a PRR for A. fumigatus that is required for proper antifungal activity of NK cells [25].Although the identity of fungal CD56 ligands remained unclear, it was known that only A. fumigatus hyphae but not resting conidia are capable to activate NK cells [44].Herein, we identified GAG as the PAMP of A. fumigatus hyphae that interacts with CD56 on NK cells, thereby activating NK cells and inducing potent secretion of chemokines and cytotoxic effector molecules, mirroring previous findings upon A. fumigatus stimulation [25,44].Notably, GAG-deficient mutants or A. fumigatus hyphae pretreated with GAG-depleting hydrolases failed to elicit NK-cell activation and secretion of treatment.(A-B) N = 6 independent donors.Repeated measures (RM) one-way analysis of variance (ANOVA) with Dunnett's post-hoc test versus Control, i.e., unstimulated NK cells (asterisks).Additionally, enzymatic pre-treatments of A. fumigatus WT and the Δagd3 mutant, respectively, were compared using RM one-way ANOVA with Dunnett's post-hoc test versus no enzymatic pre-treatment (hash signs).Columns and error bars indicate means and standard deviations, respectively.*/# p < 0.05, **/## p < 0.01, ***/### p < 0.001.https://doi.org/10.1371/journal.ppat.1012315.g005

PLOS PATHOGENS
CD56 -Aspergillus fumigatus interaction is triggered by galactosaminogalactan GAG was reported to be located at the surface of the hyphal cell wall and profusely secreted in the extracellular matrix [33] where it plays a critical role in host-pathogen interactions [30].This polysaccharide is considered a multifunctional virulence factor contributing to biofilm formation [33], host cell adhesion [32,33,45], immune evasion [27], epithelial cell damage [33], and platelet activation [46,47].Moreover, the level of cell wall GAG production is directly associated with the varying virulence within the Aspergillus genus [31].
Purified native GAG bound to CD56 and elicited pronounced NK-cell activation, ruling out that impaired NK-cellular interaction with GAG-deficient hyphae was only due to their impaired adherence [32,33,45].Moreover, we were able to specifically visualize the interaction of CD56 with hyphal cell wall structures only in GAG-producing A. fumigatus but not in GAGdeficient mutants, further corroborating that GAG is essential for hyphal-targeting by NK cells.
GAG composition appears to be critical for recognition by CD56, as A. fumigatus Δagd3, which produces GAG that lacks deacetylated GalN residues and therefore cannot stick to the hyphal surface [32], did not show binding to CD56.Given that both A. fumigatus Δagd3 and purified fully N-acetylated PGG showed no interaction with CD56, we conclude that the presence of GalN in GAG is required for the binding of CD56 to GAG.Although both native PGG and fully de-N-acetylated PGG displayed binding to CD56, the interaction with fully de-Nacetylated PGG was significantly weaker, possibly due to steric hindrance.Purified GalN oligomers but not GalNAc oligomers displayed CD56-binding-affinity and elicited NK-cell activation and chemokine secretion.This finding is consistent with prior reports that the degree of acetylation is an important modulator of the immunostimulatory versus anti-inflammatory properties of GAG [32,48].Specifically, the role of deacetylated GAG for NLRP3 inflammasome activation through translational inhibition and induction of endoplasmic reticulum stress, in turn promoting protective immunity was highlighted [49].Deacetylation of acetylated residues renders polymers cationic and introduces biologically unique properties, including adhesion to anionic surfaces such as the hyphal cell wall, plastic, and host cell membranes [32].However, even though both GAG and chitosan are de-N-acetylated polymers with cationic properties [34,35], chitosan did not exhibit immunostimulatory activity on NK cells, corroborating the specificity of the CD56/GAG interaction.
Recently, Picard et al. reported distinct effects of CD56 on primary NK-cell function and concluded that CD56 induces NK-cell degranulation, IFN-γ secretion, and morphological changes [50].Interestingly, that study did not reveal any major impact of CD56 on NK-cellular cytotoxicity.In our setting, we demonstrate that direct CD56/GAG interaction contributes to the release of antifungal effector molecules and cytotoxicity, resulting in the inhibition of hyphal growth, suggesting that GAG-activated NK cells may exert this effect by releasing perforin to reduce hyphal metabolic activity [44].Other NK-cellular effector functions that were not captured by our experiment with NK-cell supernatants, e.g., death receptor-mediated apoptosis, might also play a role.However, these effects are difficult to study, because the GAGdeficient mutants have altered growth and adhesion characteristics, making unbiased side-by-ATCC46645-GFP grown without PMNs and NK-cell supernatants (AF only, dashed line, 100%).AF grown without PMNs but with SN PGG was used as an additional control.N = 5 pairs of PMNs and NK-cell supernatants.Repeated measures (RM) 1-way analysis of variance and Dunnett's post-hoc test versus "AF only" (hash signs) and "AF + PMNs without supernatant" (asterisks) were used to determine statistical significance of the combined effect of PMNs + supernatants and the impact of the supernatant on PMN-mediated fungal growth inhibition, respectively.(C) Production of reactive oxygen species (ROS) and phagocytosis was assessed upon co-culture of PMNs with A. fumigatus in the presence of SN Control or SN PGG for 3 h (ROS production) or 1 h (phagocytosis), respectively.Results were normalized to PMNs confronted with A. fumigatus in the absence of NK-cell supernatant.N = 6 pairs of PMNs and NK-cell supernatants.Paired ratio t-test.side comparisons in co-culture experiments difficult.Our data also showed that chemokine secretion from NK cells upon GAG stimulation leads to enhanced PMN activation and engagement of PMNs into the antifungal immune response.While stimulating NK cells, GAG has been previously shown to contribute to pulmonary immunopathology by promotion of neutrophil apoptosis [29,51], induction of IL1RA secretion [48,52], and activation of the macrophage inflammasome [49].It is conceivable that NK-cellular cytotoxicity is crucial to overcome GAG-induced immunopathology.This hypothesis could explain the observation that poor NK-cell reconstitution significantly increased IPA risk in allogenic HSCT recipients, even in the absence of other cytopenia [12].
Thus far, detailed CD56 downstream signaling is largely unknown but is likely mediated by different activation kinases such as Syk, PI3K, and Erk [50].Therefore, functional studies in combination with multimodal CD56 downstream pathway analyses will be an important future direction.Furthermore, our study only focused on the NK cells-GAG interaction without considering other CD56-expressing immune cells, such as T-cell subsets, dendritic cells, and monocytes [53].CD56 on these cells might also act as a pathogen recognition receptor for A. fumigatus GAG, leading to modulation of their activation status, increased cytokine production, or altered migratory behavior.Further investigation of the crosstalk between A. fumigatus and CD56-expressing T cells and monocytes could provide insights into the broader immunomodulatory effects of this interaction.Unfortunately, no CD56 knockout mouse model is available, as no immediate CD56 homologue is present in murine NK cells.Therefore, future studies using CD56-deficient human NK cells, e.g., by means of CRISPR/Cas9 to specifically delete CD56, would be needed.Our study is subject to further limitations, including the partial lack of functional data due to the absence of an in vivo model for CD56 and the sole use of isolated NK cells without consideration of their complex interplay with other immune cell populations or tissue.Lastly, because this study focused on identifying and characterizing the fungal ligand of CD56, we did not include other fungal PAMPs and human PRRs in our analyses.
In summary, our data shed new light on the distinct A. fumigatus-NK-cell interaction by providing inaugural evidence that GAG is an activating ligand of CD56 on human NK cells.To our knowledge, in this study, we have identified the first microbial ligand recognized by this receptor.Detailed insights into the specific immunomodulatory cellular signatures and downstream pathways involved in the CD56/GAG interaction could help to develop new alternative antifungal immunotherapies involving NK cells (e.g., chimeric antigen receptor NK cells).Moreover, further insights into the specific immunoregulatory properties and virulence attributes associated with specific biochemical patterns of the GAG molecule could enable the development of new targeted anti-virulence strategies and vaccines against A. fumigatus.

Ethics statement
The processing of human peripheral venous blood from healthy adult donors was approved by the Ethics Committee of the University Hospital Wu ¨rzburg (#302/12).Healthy volunteers donating venous whole blood for isolation of PMNs gave informed written consent.

Culture of fungal strains and preparation of germ tubes
A. fumigatus strains used in this study included Af293, Δagd3 mutant [32], Δuge3 mutant [33] (kindly provided by Donald Sheppard McGill University, Montreal, Quebec, Canada), as well as American Type Culture Collection (ATCC) reference strain 46645 and ATCC46645-GFP.In addition, A. nidulans strain ATCC11267 (Leibniz Institute, DSMZ-German Collection of Microorganisms and Cell Cultures, DSM820) was used. A. fumigatus and A. nidulans strains were plated on beer wort agar (Oxoid, Wesel, Germany, PO5055A) and incubated at 35˚C until conidiophores were visible.Conidia suspensions were prepared by rinsing the plates with sterile distilled water and passed through a 20μm cell strainer (Miltenyi Biotec, Bergisch Gladbach, Germany) to eliminate residual mycelium.To generate germ tubes, 2×10 7 conidia were incubated in RPMI medium (20 mL in 50 mL tubes) at 25˚C while shaking (200 rpm) until small germ tubes were visible.Germ tubes were centrifuged at 5,000 ×g for 10 min and resuspended in RPMI medium (4×10 6 /mL).

Periodate/ Proteinase K treatment of A. fumigatus germ tubes
A. fumigatus ATCC46645 conidia were seeded in 6-well plates (2×10 5 conidia per well in 2 mL of RPMI medium) and incubated for 18 h at 30˚C.The supernatant was removed, and germ tubes were treated with 2 mL of 10 mM sodium m-periodate (Sigma Aldrich, St. Louis, MO, USA, #71859) in RPMI, or 50 μg/mL Proteinase K (Roche, Basel, Switzerland, #41227400) for 60 min in RPMI for 10 min and 60 min, respectively, or were left untreated.Thereafter, wells were washed thrice by replacing two thirds of the supernatant with HBSS buffer supplemented with 2 mM EDTA and 1% FCS.After the last washing step, the washing buffer was completely removed, and 1 mL of pre-warmed RPMI containing 10% FCS was added to the hyphae.

PLOS PATHOGENS
CD56 -Aspergillus fumigatus interaction is triggered by galactosaminogalactan Three million NK cells were added to each well in a final volume of 3 mL per well.A. fumigatus/NK-cell co-cultures were incubated for 3 h at 37˚C, 5% CO 2 and subsequently analyzed by flow cytometry.Additionally, the effect of periodate treatment was analyzed by bulk RNA-Seq.
Color development was stopped with 4% H 2 SO 4 (Sigma Aldrich, St. Louis, MO, USA, #258105-1L-PC) after 20 min.Optical density was measured at 492 nm as a surrogate of CD56 binding.Three independent experiments were performed.For the dose-dependent interaction of CD56 with GAG three technical replicates with two different batches of GAG were used.

Pulldown assay
Soluble CD56 (1 μg) was incubated with 10 μg of insoluble GAG or β-1,3-glucan (extracted from A. fumigatus) at 37˚C for 1 h.Thereafter, the contents were centrifuged to collect supernatants; pellets were washed with PBS.Both supernatants and pellets were boiled with SDS-PAGE loading buffer and subjected to SDS-PAGE (on 12% gel).

Inhibition of CD56 binding by GAG oligosaccharides
GAG was subjected to acid hydrolysis with 2 M HCl (Gibco, Thermo Fisher Scientific, Waltham, MA, USA, #H/1200/PC15), 100˚C for 3 h to obtain soluble oligosaccharides of various length (max.25-28 monosaccharide units), as previously described [29,48].PGG (10 μg/well) was coated on a 96-well microtiter plate overnight at RT.Recombinant CD56 was pre-incubated with acid-hydrolyzed GAG oligosaccharides at RT for 30 min.The mixture was centrifuged, and the supernatant was added to the PGG-coated microtiter plate and incubated at RT for 1 h.Primary (anti-CD56 antibody 1:50 in PBS + 1% BSA) and secondary (anti-mouse IgG peroxidase-conjugated, 1:1000 in PBS + 1% BSA) antibodies were added at RT for 1 h.O-phenylenediamine dihydrochloride was used as substrate to measure absorbance at 492 nm, as described above.Two independent experiments were performed.

Stimulation of PMN oxidative burst and phagocytosis by NK-cell supernatants
To generate culture supernatants, 2×10 5 NK cells were incubated with 40 μg/mL PGG for 24 h in a 200-μL volume.NK cells were centrifuged at 300 ×g for 10 min; supernatant was harvested and stored at-80˚C until further use.To assess oxidative burst, 2×10 5 isolated PMNs (2×10 6 cells/mL) were incubated with A. fumigatus ATCC46645 germ tubes (MOI 0.5) in a total of 25 μL RPMI + 10% FCS or seeded in an equivalent volume of RPMI + 10% FCS without fungal cells in a 96-well plate.Next, culture supernatant of NK cell-PGG co-cultures or NK cells cultured alone (40 μL, containing secreted metabolites of 40.000NK cells) were added.The volume of each well was adjusted to 200 μL with cell-free RPMI + 10% FCS.A combination of 5 ng/mL PMA and 0.2 μg/mL ionomycin served as positive control.ROS production was analyzed by flow cytometry, as described below.To assess phagocytosis, 2×10 5 PMNs were seeded in 96-well plate.Culture supernatant of NK cell-PGG co-cultures or NK cells cultured alone (40 μL, containing secreted metabolites of 40.000NK cells) and A. fumigatus (ATCC46645-GFP) conidia were added at an MOI of 3 for 1 h at 37˚C.The volume of each well was adjusted to 200 μL with cell-free RPMI + 10% FCS.Phagocytosis rate was determined using flow cytometry.

RNA isolation and bulk transcriptomic profiling
Total RNA from purified NK cells was isolated using QIAshredder columns (Qiagen, Hilden, Germany, Cat#79656) and the RNeasy Plus Mini Kit (Qiagen, Hilden, Germany, Cat#74136) following the manufacturer's protocol.RNA purity and concentration were tested with a NanoDrop ND-1000 spectral photometer (Thermo Fisher Scientific, Waltham, MA, USA).

PLOS PATHOGENS
The integrity of RNA was determined with a 2100 Bioanalyzer (Agilent Technologies, Waldbronn, Germany) using RNA 6000 Pico or Nano LabChip Kits (Agilent Technologies, Waldbronn, Germany) according to the manufacturer's instructions.RIN values of our samples were in the range of 7.3-10.0,indicating good quality of the RNA samples.
Library preparation was performed with the Illumina TruSeq Stranded mRNA technology, according to the manufacturer's protocol.RNA sequencing was performed by IMGM Laboratories GmbH (Martinsried, Germany) on the Illumina NextSeq 500 next-generation sequencing system with 1×75 bp single-read chemistry.Raw files are accessible under the Gene Expression Omnibus accession number GSE241020.

RNA-seq data processing and analysis
Preprocessing of raw reads, including quality control and gene abundance estimation, was performed with the GEO2RNAseq pipeline (v0.100.1)[59] in R version 3.5.1.Quality analysis was performed before and after trimming with FastQC (v0.11.7).Read-quality trimming was performed with Trimmomatic (v0.36).Adapter sequences were removed, window size trimming was performed (15 nucleotides, average Q < 25), 5 0 and 3 0 clipping was performed for any base with Q < 3, and sequences shorter than 30 nucleotides were removed.Reads were mapped against the human reference genome (GRCH 38, v109).First, the reference genome was indexed with exon information using HiSat2 (v2.1.0).Then, read alignment was performed using HiSat2 on the exon-indexed reference genome.Gene abundance estimation was performed with featureCounts (R package Rsubread, v1.34.0) in single-end mode with default parameters.MultiQC (v1.7) was used to summarize the output of FastQC, Trimmomatic, HiSat, featureCounts, and SAMtools.Count matrices were normalized using median-by-ratio normalization (MRN) as described before [60].Differential gene expression was analyzed by applying DESeq2 (v1.38.3) and adding donor origin as co-factor to the statistical design.Gene expression differences were considered significant at an adjusted p-value of �0.05.Hierarchical clustering was performed with the ward.D2 clustering method using MRN gene-abundance data with the R package ComplexHeatmap (v2.14.0).

Flow cytometry
NK-cell activation and degranulation as well as neutrophil activation were assessed by flow cytometry.Therefore, NK-cell/PMN cultures were centrifuged for 10 min at 300 ×g at RT and cell pellets were resuspended in 100 μL HBSS supplemented with 2 mM EDTA and 1% FCS.Cells were stained extracellularly with fluorescent antibodies listed in S2 Table and the fixable Viobility Live/Dead Dye (Miltenyi Biotec, Bergisch Gladbach, Germany, Cat#130-110-205).After incubation in the dark for 15 min at 4˚C, cells were washed with 2 mL HBSS supplemented with 2 mM EDTA and 1% FCS, resuspended in 150 μL of 4% paraformaldehyde (Sigma-Aldrich, St. Louis, MO, USA, #47608), and incubated for 30 min at RT. Intracellular staining was performed using the Inside Stain Kit (Miltenyi Biotec, Bergisch Gladbach, Germany, Cat#130-090-477).After extracellular staining, cells were washed with 2 mL HBSS supplemented with 2 mM EDTA and 1% FCS, resuspended in 300 μL of Inside Fix solution (150 μL Inside Fix and 150μL HBSS + EDTA), and incubated for 20 min at RT.After washing, 100 μL of intracellular staining mix in Inside Perm was added and cells were incubated for 10 min at RT in the dark.

Fluorescence microscopy of NK-cell/A. fumigatus co-cultures
A. fumigatus Af293 and Δagd3, (5×10 4 conidia per well in colorless RPMI) were seeded on laminin-coated (20 μg/mL) 8-well coverslips and incubated overnight at 33˚C to enable germination, as described above.NK cells were added to the germ tubes at an MOI of 0.5 (1×10 5 cells per well) and incubated for 4 h at 37˚C, 5% CO 2 .After incubation, samples were fixed with 1% formaldehyde for 7 min, washed once with PBS, and blocked with 2.5% BSA in PBS for 30 min.Samples were stained immediately after fixation with Alexa Fluor 647-labeled mouse-anti-human CD56 (10 μg/mL) for 1 h and washed with PBS.CLSM images were recorded with a LSM700 system (Carl Zeiss, Jena, Germany) using a plan-apochromat 63 ×/1.40 oil immersion objective.

IncuCyte imaging
To generate culture supernatants for IncuCyte experiments, 2×10 5 NK cells were incubated with 40 μg/mL PGG for 24 h in a 200-μL volume.NK cells were centrifuged at 300 ×g for 10 min; supernatant was harvested and stored at-80˚C until further use.For IncuCyte experiments, A. fumigatus spores (ATCC46645-GFP) were diluted in RPMI + 10% FCS to a final concentration of 3×10 3 /mL.Fifty microliters (150 conidia/well) were dispensed per well of a 96-well flat-bottom plate.To test the direct antifungal effect of PGG-pulsed NK cells, culture supernatant of NK cell-PGG co-cultures or NK cells cultured alone (37.5 μL, containing secreted metabolites of 37,500 NK cells) was added to the wells.The volume of each well was adjusted to 150 μL with cell-free RPMI + 10% FCS.To evaluate the capacity of PGG-stimulated NK cells to enhance fungal growth inhibition by PMNs, 1.5×10 4 PMNs (effector/target ratio, 100) in 85 μL RPMI + 10% FCS were added to the wells.In addition, culture supernatant of NK cell-PGG co-cultures or NK cell cultured alone (15 μL, containing secreted metabolites of 15,000 NK cells) was added to the wells.The volume of each well was adjusted to 150 μL with cell-free RPMI + 10% FCS.In addition to biological replicates (n = 5), each condition was tested in technical duplicates.IncuCyte microscopy and NeuroTrack-based image analysis (IncuCyte Zoom NeuroTrack software module) was performed according to Wurster et al. [39].In brief, well plates were imaged hourly in the IncuCyte Zoom HD/2CLR time-lapse microscopy system (Sartorius, Go ¨ttingen, Germany) equipped with an IncuCyte Zoom 10× Plan Fluor objective (Sartorius, Go ¨ttingen, Germany) for a period of 16 h.Acquisition time for the green channel was 400 ms.The following parameters were used for NeuroTrack analysis: neurite coarse sensitivity, 5; neurite fine sensitivity, 0.25; neurite width, 4 μm.Neurite length [mm/mm 2 ] and numbers of branch points [1/mm 2 ]) were compared to an "A.fumigatus only" control without NK-cell supernatants and/or PMNs.

Statistical analyses
Microsoft Excel 365, GraphPad Prism version 9, and R version 3.5.1 were used for data compilation, analysis, and visualization.Significance testing was performed using one-way analysis of variance (ANOVA), repeated measures one-way ANOVA, and paired t-test for unmatched multi-group comparisons, matched multi-group comparisons, and matched two-group comparisons, respectively.For multi-group comparisons, Dunnett's or Tukey's post-hoc tests were used for experiments with a single control group or analysis of all pairwise comparisons, respectively.Significance tests are specified in the figure legends.(Adjusted) p-values < 0.05 were considered significant.

Fig 4 .
Fig 4. Fully N-acetylated PGG and GalNAc oligomers fail to interact with CD56 and elicit no NK-cell activation.(A) CD56 binding to fully acetylated (aPGG), fully deacetylated (dePGG), and native galactosaminogalactan (PGG), was determined by enzyme-linked immunosorbent assay (ELISA).Additional conditions with incomplete ELISA setup were included to preclude unspecific binding of the secondary antibody or anti-CD56 to the carbohydrates.N = 3 technical replicates.One-way analysis of variance (ANOVA) with Tukey's post-hoc test was performed for each assay setup.(B) CD56, Fig 5A) and triggered strong production of the cytotoxic effector molecules perforin and granzyme B (Fig 5B).

Fig 5 .
Fig 5.The complete, GalN-rich galactosaminogalactan molecule is required for interaction of CD56 with Aspergillus fumigatus.(A) CD56 and CD69 expression of unstimulated NK cells (Control) and A. fumigatusstimulated NK cells (6 h) depending on the fungal strain (wild type [WT] Af293 or isogenic mutants with defective GAG biosynthesis) and its enzymatic pre-treatment.(B) Secretion of granzyme B and perforin by unstimulated NK cells (Control) and A. fumigatus-stimulated NK cells depending (6 h) on the fungal strain and its enzymatic pre-

Fig 6 .
Fig 6.PGG-pulsed NK-cell supernatants confer a direct antifungal effect and engage PMNs into the antifungal immune response.(A) Growth and morphogenesis of ATCC46645-GFP germ tubes in the presence of supernatants from unstimulated NK cells (SN Control) or PGG-stimulated NK cells (SN PGG) was quantified using the IncuCyte NeuroTrack assay and compared to ATCC46645-GFP grown without NK-cell supernatants (AF only).Left panel: Kinetics of hyphal length during transition from germ tubes to hyphae (6-16 h of culture).Center and right panel: Hyphal length and branch point numbers after 16 h of culture, normalized to the "AF only" control (dashed line, 100%).The assay was performed using NK-cell supernatants from n = 4 independent donors.Paired ratio t-test.In addition, images from a representative donor are shown.Green overlays indicate hyphal detection by the NeuroTrack algorithm.(B) Growth and morphogenesis of ATCC46645-GFP germ tubes in the presence of polymorphonuclear neutrophils (PMNs), with or without addition of SN Control or SN PGG, was quantified using the IncuCyte NeuroTrack assay.Hyphal length and branch point numbers after 16 h of culture were normalized to (A-C) Columns and error bars indicate means and standard deviations, respectively.* p < 0.05, ** p < 0.01, *** p < 0.001.https://doi.org/10.1371/journal.ppat.1012315.g006 Cells were washed with 750 μL Inside Perm and resuspended in 150 μL of HBSS with EDTA.Data were acquired on a FACSCalibur (Treestar/Becton & Dickinson, BD), Cytoflex flow cytometer (Beckman Coulter, Brea, CA, USA) or MACS Quant 10 flow cytometer (Miltenyi Biotec, Bergisch Gladbach, Germany).Downstream data analysis was performed with FlowJo10 (Treestar/Becton & Dickinson, BD) or Kaluza v.2.1 (Beckman Coulter, Brea, CA, USA).Gating strategies and representative raw data are shown in S2, S3, S4 and S5 Figs.

S4
Data.Detailed information about generated raw dataset shown in Fig 4. (XLSX) S5 Data.Detailed information about raw dataset used to generate Fig 5. (XLSX) S6 Data.Detailed information about the generated primary material of Fig 6. (XLSX) S7 Data.Detailed information about raw dataset used to generate S1 Fig. (XLSX)