Functional surface proteomic profiling reveals the host heat-shock protein A8 as a mediator of Lichtheimia corymbifera recognition by murine alveolar macrophages.

Mucormycosis is an emergent, fatal fungal infection of humans and warm-blooded animals caused by species of the order Mucorales. Immune cells of the innate immune system serve as the first line of defense against inhaled spores. Alveolar macrophages were challenged with the mucoralean fungus Lichtheimia corymbifera and subjected to biotinylation and streptavidin enrichment procedures followed by LC-MS/MS analyses. A total of 28 host proteins enriched for binding to macrophage-L. corymbifera interaction. Among those the HSP70-family protein Hspa8 was found to be predominantly responsive to living and heat-killed spores of a virulent and an attenuated strain of L. corymbifera. Confocal scanning laser microscopy of infected macrophages revealed colocalisation of Hspa8 with phagocytosed spores of L. corymbifera. The amount of detectable Hspa8 was dependent on the multiplicity of infection. Incubation of alveolar macrophages with an anti-Hspa8 antibody prior to infection reduced their capability to phagocytose spores of L. corymbifera. In contrast, anti-Hspa8 antibodies did not abrogate the phagocytosis of Aspergillus fumigatus conidia by macrophages. These results suggest an important contribution of the heat-shock family protein Hspa8 in the recognition of spores of the mucoralean fungus L. corymbifera by host alveolar macrophages and define a potential immunomodulatory therapeutic target. This article is protected by copyright. All rights reserved.


Introduction
Mucormycosis is a potentially life-threatening fungal infection particularly dangerous for patients suffering from immune deficiency, e.g. during long-term corticosteroid treatment, trauma, and solid-organ transplantation or uncontrolled diabetes (Skiada et al., 2013). The incidence and mortality rate of mucormycosis have notably increased during the last decades, especially in developing countries (Hassan and Voigt, 2019). The treatment of mucormycosis is hindered by the lack of appropriate diagnostic tools, the resistance of mucoralean fungi to most available antifungal agents (Ma et al., 2009), and the lack of a mechanistic understanding of the interaction between the immune system and mucoralean fungi (Skiada et al., 2018). Some recent studies, however, have identified factors that are specifically involved in mechanisms of host-pathogen interactions targeting these pathogenic fungi, e.g. eumelanin found in human skin and hair, the spore coat protein CotH, and the multifunctional regulator calcineurin were all proven to be associated with the pathogenicity of Rhizopus oryzae and Mucor circinelloides (Gebremariam et al., 2014;Andrianaki et al., 2018;Hassan and Voigt, 2019). On the host side, pattern recognition receptors are proteins that mediate the antifungal immune response from the surface of immune cells where they are involved in the identification of Mucorales-related pathogen-associated molecular patterns (PAMPs) and damage-associated molecular patterns (DAMPs) (Suresh and Mosser, 2013). Liu et al. (2010) identified the glucose-regulated protein 78 (GRP78) receptor (synonymous: Hspa5), which is a member of the HSP70 protein family on the surface of endothelial cells that mediates adherence and invasion of R. oryzae germlings (Liu et al., 2010). However, because endothelial cells are not professional phagocytes, the identification of surface proteins of innate immune cells important for recognition of mucoralean fungi is of great interest to understand pathogenicity.
Pulmonary mucormycosis is considered the most common type of mucormycosis, with the major route of infection through inhalation of mucoralean fungal spores (Fernandez et al., 2013). Alveolar macrophages (AM), dendritic cells (DC) and polymorphonuclear cells of the innate immune system remove inhaled spores (Martin and Frevert, 2005). AM represents the major fraction of leukocytes in the human lung, comprising approximately 95% of the immune cells in lung tissue (Martin and Frevert, 2005). AM produces compounds including cytokines, chemokines and calreticulin that help to activate host immunity and kill fungal spores (Li et al., 2013;Krysko et al., 2018). Several studies have reported the interaction of mucoralean fungi with macrophages (Gebremariam et al., 2014;Kraibooj et al., 2014;Andrianaki et al., 2018;Hassan and Voigt, 2019). Spores of different mucoralean fungi showed variation in their phagocytosis rates (Waldorf et al., 1984;Li et al., 2011;. Twentysix species of the order Mucorales have been found to cause mucormycosis, such as species of the genera Lichtheimia, Rhizopus, Mucor and Cunninghamella (Ribes et al., 2000;de Hoog et al., 2019). Lichtheimia species rank as the second and third leading cause of mucormycosis in Europe and the United States of America respectively (Ribes et al., 2000). On the other hand, mucoralean fungi can also cause an allergic disorder that is known as farmer's lung disease. Farmer's lung disease is mostly caused by Lichtheimia species during frequent inhalation of Lichtheimia spores from agricultural waste.
The spores reach the lung alveoli, where they can cause hypersensitivity pneumonitis (Bellanger et al., 2010). Despite this clinical importance, little is known about the molecular mechanisms of detection of L. corymbifera spores by the immune system. In our current study, we discovered proteins on the surface of AM that recognize mucoralean fungi and thus are of importance for defence against these pathogens. The identified proteins might also provide promising targets for improving therapy against the invasive infections.

Results
Heat-shock proteins are the predominant AM surface proteins bound to L. corymbifera The modification of surface proteins by biotinylation has recently emerged as an elegant biochemical technique to specifically isolate and quantify cell surface proteins of immune cells that bind to pathogens (Liu et al., 2010). After labelling surface proteins of macrophage cells with EZ-Link Sulfo-NHS-SS-biotin, macrophages were lysed and the membrane-associated proteins were isolated by ultracentrifugation. Then, the membrane protein fraction was incubated with dormant spores of L. corymbifera, followed by subsequent isolation of the biotin-tagged proteins bound to the spore surface by streptavidin pull-down. Macrophage surface proteins bound either to the living or heat-inactivated spores of a virulent or attenuated strain of L. corymbifera were then identified by LC-MS/MS (-Tables S1 and S2) (Schwartze et al., 2012). Of note, heatinactivated spores of both strains were included in our study as controls. Three biological replicates were analysed for each treatment. The proteins bound to the surface of L. corymbifera spores that were detected in at least two biological replicates with peptide spectrum match (PSM) values higher than or equal to 15 were included for further evaluation. A total of 573 macrophage proteins were observed across all conditions to associate with living and heat-inactivated L. corymbifera spores after confrontation. Interestingly, we found considerable differences in the amount of macrophage proteins binding to the virulent and attenuated L. corymbifera strains, FSU 9682 and FSU 10164 respectively. For the virulent strain, a total of 424 macrophage surface proteins were detected during confrontation with living spores, while only 152 proteins were found during confrontation with heat-killed spores. In contrast, 34 macrophage proteins, a significantly reduced number were found to interact with living spores of the attenuated strain of L. corymbifera, while 446 macrophage surface proteins were detected to bind after exposure to heat-killed spores.
As shown in the Venn diagram in Fig. 1A, there are also striking differences between the number of macrophage proteins binding to living and heat-killed spores of each strain (Tables S3-S6). Thirty-three macrophage proteins were detected that were binding to living spores of both the virulent and the attenuated strain after confrontation. On the contrary, 143 macrophage proteins were identified that bound to either heat-killed spores of the virulent or the attenuated strains of L. corymbifera. A total of 151 macrophage proteins were shared that bound to living and heat-killed spores of the virulent L. corymbifera strain, while only 33 macrophage proteins bound to the living and heat-killed spores of the attenuated strain of L. corymbifera. We identified 28 macrophage membrane-associated proteins that were identified in all tested L. corymbifera strains ( Fig. 1A; Table 1). These proteins belonged to various groups, such as actin cytoskeleton proteins, metabolic proteins, calciumbinding chaperones, elongation factors, ATP binding proteins, and, most prominently, HSPs, such as Hspa90, Hspa5 and Hspa8. In particular, HSPs dominated the population of macrophage surface proteins that adhered to the spore surface (Table 1).
In general, the level of these 28 macrophage proteins that bound to all L. corymbifera samples did not change significantly between the virulent and attenuated living and heat-killed spores (Table 2). However, the glycolytic enzyme phosphoglycerate kinase 1 (Pgk1) and the 14-3-3 family protein YWHAZ were significantly enriched during the confrontation of macrophages with living spores of the virulent strain (p = 0.001 and 0.038 respectively) ( Table 2). The comparison of proteins in living versus heat-killed spores of the attenuated strain revealed the dominance of HSPs, glycolytic enzymes, moonlighting proteins with multiple known functions, mitochondrial membrane ATP synthase, thiol-specific peroxidase, phosphoglucan phosphatase DSP4 and annexin A3 in interaction with the spore (Table 2).
In turn, Hspa5, Pgk1, Eno1, Pkm, Atp5a1, Hmgb1, Eef1a1, Hsp90b1 and Anxa3 were significantly lower in abundance on macrophage surfaces when confronted with living spores of the virulent strain compared with the attenuated spores (Table 3). Eef1a1 was less abundant during confrontation with living spores compared with heat-killed spores of the virulent strain. On the contrary, no significant difference was observed in the macrophage proteins that exhibit lower binding in response to the interaction with living versus heat-killed spores of the attenuated strain. However, Pgk1 and Hspa5 proteins were less abundant during confrontation with heat-killed spores of the virulent strain compared with heatkilled spores of the attenuated strain of L. corymbifera (Table 3).
We repeated these experiments without biotinylating the macrophage surface as a control to detect macrophage proteins not present on the surface that may nonspecifically bind to streptavidin beads. Only nine out of a total of 573 macrophage surface proteins were found to bind non-specifically (Fig. 1B). The proteins identified as 'unspecific' refer to keratins and proteolytic enzymes, as well as glycolytic proteins (p < 0.05 among all conditions, GAPDH,Krt10,Krt15,Prdx1,Dsp,Krt78,Krt90,Krt6b and Try10). These data confirmed that the detected proteins of macrophages are likely specifically binding to the spores of L. corymbifera strains.
Gene ontology (GO) analysis revealed the enrichment of a variety of different functions of macrophage surface proteins binding to living spores of the virulent strain compared with the attenuated strain. For example, proteins involved in cell adhesion, homeostatic and metabolic processes, as well as in response to chemical stress were all enriched on the spores of the virulent strain ( Fig. S1A; Table S7). In contrast, only the carbohydrate catabolic pathway was found to be less abundant in binding to the virulent compared with the attenuated strain ( Fig. S1B; Table S8). Analysis of macrophage proteins binding to living spores compared with heat-killed spores of the virulent strain revealed protein sets involved in intracellular transport (e.g. Golgi vesicle transport) ( Fig. S1C; Table S9), whereas the comparison of living versus heatkilled spores of the attenuated strain revealed sets of proteins that mediate cellular localization ( Fig. S1D; Table S10).
Hspa8 is the most abundant macrophage surface protein that binds to L. corymbifera We next determined the most abundant macrophage proteins bound to spores of L. corymbifera using an alternative method. We performed Western blot analysis of biotinylated macrophage protein extracts confronted with L. corymbifera Twenty eight MH-S proteins are divided into seven groups based on the GO molecular functions comprising ATP binding, calcium-binding proteins, cytoskeletal proteins, heat-shock proteins, metabolism, nucleotide binding and protein binding.
spores using an anti-biotin antibody and subsequently identified a single prominent band of approximately 70 kDa under all conditions tested ( Fig. 2A). MALDI-TOF/TOF analysis of the excised band revealed oligopeptides with an unequivocal sequence matching to the heat-shock 70 kDa (Hsp70) family A (Hsp70) protein 8 (Hspa8; synonymous: heat-shock cognate 71 kDa protein, lipopolysaccharideassociated protein 1, N-myristoyltransferase inhibitor protein 71, constitutive heat-shock protein 70, epididymis luminal protein, epididymis secretory sperm-binding protein Li 72p, Fig. S2). These results agreed with the LC-MS/MS data (Table 1), which indicated Hspa8 as one of the 28 major macrophage protein candidates bound to spores of L. corymbifera. This result prompted us to further characterize the role of Hspa8 in the binding of macrophages to spores of L. corymbifera.

Hspa8 is abundant on the surface of macrophages
The observed abundance of the Hspa8 protein on the surface of macrophages was confirmed by flow cytometry analysis using a commercially available anti-Hspa8 antibody ( Fig. 2B and C). Subsequent statistical analysis revealed significant differences between macrophage cells stained with the anti-Hspa8 antibody compared with control samples, consisting of macrophage cells without antibody or macrophage cells stained with the secondary antibody only. Furthermore, flow cytometry results also confirmed that host Hspa8 was not present on the surface of living spores of both the virulent and the attenuated strain of L. corymbifera (Fig. 3). The flow cytometry results were also supported by confocal microscopy, which showed positive staining of macrophage cells with the anti-Hspa8 antibody ( Fig. 4A-F). In addition, the permeabilization of macrophages led to the colocalization Table 3. Summary of statistical significance difference in downregulation of 28 MH-S proteins during confrontation with spores in various conditions (live and heat-killed) of L. corymbifera strains (virulent and attenuated).

p-values
No. of Hspa8 with engulfed spores of L. corymbifera (virulent and attenuated strain) ( Fig. 4G-L). These data suggest that Hspa8 can be exclusively found on macrophage cells but not on L. corymbifera spores.
Hspa8 cell surface expression is dependent on the infectious dose of fungal spores To determine whether a multiplicity of infection (MOI) of 1, 3 and 5 influenced the abundance of Hspa8 on the surface of macrophage cells, we conducted flow cytometry analyses of infected macrophages, which revealed a counterintuitive relationship between MOI and the level of Hspa8 positive signals. The higher the MOI, the lower the fluorescence intensity of Hspa8 on the surface of macrophage cells (Fig. 5A and B). Co-incubation of macrophage with living spores of the virulent L. corymbifera strain at an MOI of 1 led to some background fluorescence, which was comparable to macrophages that had not been confronted with spores. However, the positive fluorescence signal displaying the presence of Hspa8 was significantly reduced by more than 60% when macrophages were exposed to spores of L. corymbifera at an MOI of 3 or 5. This observation may be explained by the translocation of Hspa8 from the membrane to the cytoplasm at higher MOIs as evidenced by our experiments with permeabilized macrophages (Fig. 5). Another explanation for the reduction of fluorescence intensity at higher MOI is the possibility that the interaction between AM-Hspa8 and the spores could block the recognition by the anti-Hspa8 antibody (competition between the targeted PAMP and the antibody to detect Hspa8).
Inhibition of phagocytosis of spores from L. corymbifera by antibody-mediated blocking of Hspa8 In order to determine the role of Hspa8 in recognition and endocytosis of L. corymbifera spores by macrophage cells, we confronted macrophage cells with anti-Hspa8 polyclonal antibodies and determined the phagocytic index ( Fig. 6). Blocking of Hspa8 by antibodies revealed a decrease of phagocytosis of spores from both virulent and attenuated L. corymbifera strains by macrophages pretreated with anti-Hspa8 antibody compared with the rabbit IgG isotype control antibody. The anti-Hspa8 antibody inhibited the phagocytosis of the spores from the virulent and the attenuated strains by 50% and 25% respectively ( Fig. 6A and B). Interestingly, blocking of macrophage cells with anti-Hspa8 antibody did not influence the capability of macrophages to phagocytose A. fumigatus conidia (Fig. 6C). These data confirmed that Hspa8 is important for the recognition and the phagocytosis of L. corymbifera, but is dispensable for the ascomycetous fungus A. fumigatus.
In addition, Hspa8 did not affect the adherence of L. corymbifera spores or A. fumigatus conidia to macrophages (data not shown). Ultimately, these results indicate that Hspa8 serves an important function as a receptor for L. corymbifera on the surface of AMs.

Discussion
AMs belong to the first line of immune defence in lung infections. The main goal of this study was to determine surface proteins of AM that are involved in binding the surface of the spores of L. corymbifera. Previous studies showed that macrophages have various proteins or receptors on their surface such as the melanin-sensing C-type lectin receptor (MelLec). MelLec is expressed on the surface of endothelial and myeloid cells and recognizes dihydroxynaphthalene-melanin that covers the surface of spores of A. fumigatus (Stappers et al., 2018). Other host surface proteins are able to bind to specific pathogen surface proteins as well. For example, integrin α5β1 is a receptor on the surface of endothelial cells, which binds to the A. fumigatus CalA surface protein (Liu et al., 2016). Moreover, C. albicans is recognized by N-cadherin expressed on endothelial cells (Phan et al., 2005) and is also detected by the ephrin type-A receptor 2 (EphA2 receptor) on epithelial cells (Swidergall et al., 2018). For mucoralean fungi, a recent study identified the GRP78 as the receptor on the surface of endothelial cells that recognizes germlings of R. oryzae (Liu et al., 2010). It should be added that the GRP78 receptor specifically binds to mucoralean fungi but not to other invasive fungal ascomycetous pathogens such as A. fumigatus (Liu et al., 2010). In L. corymbifera, two strains were identified as virulent and attenuated in invertebrate, avian and murine infection models (Schwartze et al., 2012;Schulze et al., 2017). Both strains differ significantly in the phagocytosis by murine AMs, which imply substantial changes in the spore cell wall (Kraibooj et al., 2014). In order to address the architecture of the cell wall in the context of virulence, we investigate the surface proteome of the macrophage in response to resting spores in a comparative approach utilizing both strains. In total, 28 spore-binding surface proteins of macrophages were identified. Several of these proteins are particularly interesting. The ATP synthase subunit β protein (ATP5B) has been previously described to be important for the host cell to recognize the hepatitis E virus (Nakamoto et al., 2016). The alpha enolase protein (Eno1) is a key player in glycolysis but has a second function as a receptor for plasminogen that inhibits the complement cascade (Agarwal et al., 2012). Furthermore, it was shown that Eno1 plays a key role in immune defence against infectious disease through upregulation in DCs during Chlamydia infection (Ryans et al., 2017). The largest fraction of proteins we detected on the surface of macrophage cells during their response to L. corymbifera spores belonged to the group of heat-shock proteins (HSPs). HSPs are known to play a role in folding/unfolding of proteins, transport of other proteins, cell signalling, cell shielding against stress (extracellular stress induced by chemical compounds or intracellular stress during apoptosis), cardioprotection and neuroprotection (Malyshev and Malysheva, 1998;Trivedi, 2007). Fungal HSPs are distributed throughout the entire cell (cell membrane, cytosol and endoplasmic reticulum) and have various functions, including protein folding, cell cycle processes, heme-oxygenase, cytoskeleton formation, hyphal formation, virulence, and glycogen and glycerol haemostasis as reviewed previously (Tiwari et al., 2015). Fungal HSPs also have a role in the resistance to antifungal compounds, as shown by Hsp90, which enhances the resistance of Aspergillus and Candida species to azoles and echinocandins (Robbins et al., 2017). Moreover, Hsp90 is considered as a promising therapeutic target against Trichophyton rubrum (Neves-da-Rocha et al., 2019). Overexpression of Hsp12p in C. albicans leads to increased susceptibility to antifungal agents such as fluconazole, ketoconazole and itraconazole (Fu et al., 2012). Other HSPs such as Hsp104, Hsp70 and Hsp21 play a role as antifungal resistance in C. albicans (Gong et al., 2017). Moreover, HSPs are a major constituent of DAMPs recognized by the immune system during injury or pathogen invasion (Chen and Nuñez, 2011). As A B Fig 5. Hspa8 protein on the surface of macrophage cells is dependent on the MOI. A: Flow cytometry shows the signal of Hspa8 protein on the surface of macrophage cells that were confronted with different MOIs. B: Statistical analysis of three independent biological replicates of flow cytometry data. 'ns', no significant difference (p ≥ 0.05). Asterisks show significant differences at p < 0.05, **p < 0.01 and ***p < 0.001. Images are representative triplicate experiments. [Color figure can be viewed at wileyonlinelibrary.com] mentioned above, the Hspa5 protein, also known as GRP78, is a ligand on the surface of endothelial cells that recognize germlings of R. oryzae (Liu et al., 2010). The Hspa9 protein also known as mortalin plays a key role in clathrin-mediated endocytosis and in cell cycle organization in cancer cells (Gao et al., 2017), which suggests a role for Hspa9 in endocytosis of L. corymbifera spores by macrophages. The Hsp90aa1 protein induces autophagy and serves as a receptor on the surface of chicken embryo fibroblasts during interaction with avibirnavirus VP2 (Hu et al., 2015). The Hsp60 protein is important for apoptosis, serves as a target for therapeutics against cancer cells, and protects the host cells from Hepatitis B virus and Human immunodeficiency virus, giving further evidence for HSPs as receptors for pathogens (Wyzewski et al., 2018).
In this study, we found that the most abundant protein on the surface of macrophages that recognizes L. corymbifera spores is the heat-shock protein Hspa8. Hspa8/HSC70 is a cognate protein of the HSP70 family whose members mainly localize to the endoplasmic reticulum. They play a powerful role in immune response and autophagy, consistent with our data for L. corymbifera (Stricher et al., 2013). Bacterial Hsp70 is present on the surface of Mycobacterium tuberculosis and plays a role in the interaction with immune cells by activation of the NF-κB pathway and by the secretion of pro-inflammatory compounds in the macrophages through TLR4 (Bulut et al., 2005). The expression of Hsp70 was altered on the surface of AM in patients suffering from tuberculosis compared with healthy donors (Wang et al., 2017;Ziegler et al., 2017). Further support for HSPs playing a prominent role for infection is the finding that Hsp70 levels were significantly higher in AM and epithelial cells in the lungs of patients suffering from chronic asthma and bronchitis (Vignola et al., 1995). Furthermore, hypoxia stimulates the expression of Hsp70 on the surface of macrophages and subsequently reduces the capability of Leishmania amazonensis to invade macrophages (Degrossoli et al., 2004).
Hspa5 (GRP78) shares high similarity (about 65%) at the amino acid level with Hspa8, implying that both proteins have similar functions for recognition of fungal pathogens. Together with previous observations, our findings indicate that Hspa8 plays a major player in the recognition of L. corymbifera. We also observed an MOI-dependent response for Hspa8 to L. corymbifera spores. This result is reminiscent of a previous finding showing that the abundance of the CD56 receptor on the surface of Natural Killer cells was influenced by the MOI of A. fumigatus (Wang et al., 2017;Ziegler et al., 2017). Similar to previous results for GRP78 (Liu et al., 2010), we found that the Hspa8 protein colocalizes on the surface of phagocytosed spores of L. corymbifera.
To assign a potential function to Hspa8 in the recognition of spores, we analysed its role for phagocytosis. After treatment with anti-Hspa8 blocking antibodies, we in fact observed an inhibition of phagocytosis of spores by macrophages. Recent studies demonstrated that HSC70 proteins play a role in the formation of clathrin coats on intracellular vesicles (Hannan et al., 1998), an important feature of clathrin-mediated endocytosis (Brodin et al., 2000). The importance of HSC70 proteins in clathrin-mediated endocytosis offers a potential explanation for our observed inhibition of phagocytosis of the spores of L. corymbifera after blocking of Hspa8 on the surface of AM by the anti-Hspa8 antibody. The specificity of the interaction of Hspa8 with macrophages was further demonstrated by the finding that blocking of Hspa8 on the surface of AMs did not influence the phagocytosis of A. fumigatus spores. It is also conceivable that Hspa8 binds to other proteins to promote fungal recognition, as observed for the ubiquitin-like protein of RAW 264.7 macrophage-like cells (Notsu et al., 2016). Moreover, HSPs are well-known alarmins that induce the formation of proinflammatory cytokines by innate immune cells during injury or invasion of macrophages and neutrophils by pathogens (Yang et al., 2017). Consequently, Hspa8 may also possess additional roles beyond the initial recognition of fungal spores.

Conclusions
The proteomics data collected here provide a rich resource for the further study of host fungal pathogen interaction. Here, based on a comprehensive proteomics analysis we provide evidence for the important role of the host Hspa8 in the recognition of L. corymbifera by AM. The data obtained enhance our understanding of L. corymbifera pathogenesis and identify a novel interacting protein that might function as a so far unknown receptor.

Experimental design and statistical rationale
In order to identify macrophage surface proteins that are bound to the surface of L. corymbifera spores, we biotinylated the surface of macrophages and co-incubated them with either vital or heat-killed spores of both virulent and attenuated L. corymbifera strains. After enrichment of biotinylated proteins with streptavidin beads, we performed an LC-MS/MS-based proteome analysis by using three independent biological replicates for each condition. The experiments were repeated without biotinylation of the macrophage surface as a control in order to identify and to subtract unspecifically bound proteins during the streptavidin clean-up step. In total, 24 samples (three biological replicates of virulent-heat-killed, virulent-vital, attenuated-heat-killed, and attenuated-vital samples both with and without macrophage surface biotinylation) were measured by LC-MS/MS based on every two analytical replicates (two LC-MS/MS runs per sample). GO term enrichment was performed by g: Profiler due to several advantages such as the feasibility of data from the Ensembl database. For macrophage and fungal proteins, the significance was measured through the Bonferroni method as it is a conservative method in the presence of a large number of samples and does not consider data dependencies.

Fungal strains and culture conditions
Two strains of L. corymbifera were previously reported to be virulent (JMRC:FSU:09682) and attenuated (JMRC: FSU:10164) in invertebrate, avian and mammalian infection models (Schwartze et al., 2012;Schwartze et al., 2014a;Schulze et al., 2017). Cultivation was performed on malt extract agar (40 g/l of malt extract, 4 g/l of yeast extract and 15 g/l agar) for 7 days at 37 C. All chemicals were purchased from Sigma-Aldrich unless otherwise noted. Plates were washed with phosphate buffer saline solution (PBS; 3 M NaCl, 0.5 M Na 2 HPO 4 , 0.27 M KCl, 0.5 M KH 2 PO 4 at pH 7) to harvest the spores. Spores were filtered through a 40-μm cell strainer (EASYstrainer, Greiner BIO-ONE) to remove fungal hyphae or any other cellular debris. The number of spores was counted using a haemocytometer (Neubauer chamber). Heat-killed spores were obtained by exposing the spores for 1 h to 99 C with shaking in a ThermoMixer (Eppendorf) at 500 rpm.
A wild-type strain of A. fumigatus (ATCC 46645) was cultivated for 5 days on Aspergillus minimal medium as described previously .

Extraction of membrane and surface proteins from AM by biotinylation
Cells of AM were grown to 70%-90% confluency. AMs were infected with spores at a MOI of 0.1, i.e. 10 7 spores (living or heat-killed) per 10 8 AM, in accordance with previously published protocols (Liu et al., 2010). The cells were kept on ice for 30 min to ensure direct contact between spores and macrophage cells and subsequently incubated for 30 min at 37 C in a 5% (v/v) CO 2 incubator. The supernatant was removed and the cells were washed three times with PBS to remove excessive spores. The surface of the cells was biotinylated using 0.5 mg/ml EZ-Link Sulfo-NHS-SS-Biotin (Thermo Fisher Scientific) in PBS supplemented with Ca 2+ and Mg 2+ (0.884 and 0.492 mM respectively; PBS-Ca 2+ /Mg 2+ ) (Roti®-CELL DPBS) and incubated for 12 min at 37 C in a 5% (v/v) CO 2 incubator. The biotinylated solution was removed and the cells were washed three times with PBS-Ca 2+ /Mg 2+ . AMs were scraped in PBS-Ca 2+ /Mg 2+ and centrifuged with 100g for 5 min at 4 C. The supernatant was discarded and the macrophage cells were lysed by incubation in 58 mg/ml n-octyl-β-Dglucopyranoside (Pan Reac AppliChem) on ice for 20 min in the presence of a protease inhibitor cocktail (Sigma-Aldrich). The macrophage surface proteins were collected by ultracentrifugation at 100 000g for 1 h at 4 C. The concentration of the macrophage proteins was measured using Bradford reagent (Bio-Rad) according to the manufacturer's instructions.

Isolation of macrophage surface proteins interacting with L. corymbifera spores
We followed the method previously published (Liu et al., 2010) with the following modifications: 250 μg of macrophage surface proteins were incubated with 8 × 10 8 spores at 37 C for 1 h in the presence of a protease inhibitor cocktail (Sigma-Aldrich). The spores were collected by centrifugation at 15 000g for 5 min and subsequently washed thrice with PBS-Ca 2+ /Mg 2 to remove unbound proteins. The spores were incubated with elution buffer (25 mM glycine and 1% (w/v) SDS, pH 2) for 5 min at room temperature and collected by centrifugation with 6000g for 5 min. The elution step was repeated twice.

Streptavidin enrichment of biotinylated proteins
The pooled protein solutions containing biotinylated macrophage proteins were enriched by streptavidin particles to deplete intracellular and non-biotinylated surface proteins (Rybak et al., 2004). Briefly, Roti-MagBeads Streptavidin (ROTH, HP57.1) were washed three times with buffer A (1% (v/v) NP-40, 0.1% (w/v) SDS in PBS), then added to the bound protein solution on ice for 2 h. Unbound proteins were removed by collecting streptavidin particles with a magnet and washed three times with buffer A, two times with buffer B (0.4 M NaCl in buffer A) and once with 50 mM Tris-HCl (pH 7.5). The captured proteins were released from the streptavidin beads by treatment with 400-μl of a solution containing 5% (v/v) 2-mercaptoethanol in PBS for 30 min at 30 C. This step was repeated twice. The eluted proteins were pooled, enriched by precipitation with TCA (Sigma-Aldrich) to a final concentration of 10% (w/v) for 30 min on ice and collected by centrifugation at 10 000g for 5 min. The supernatant was discarded and the pellet was washed once with 1 ml of ethanol-ether (1:1). The centrifugation step was repeated and the pellet was air dried. The precipitate was dissolved in 200 μl of 50 mM NH 4 HCO 3 and sonicated for 15 min at room temperature. The dissolved precipitate was denatured for 10 min at 95 C and equilibrated on ice. The reduction step was carried out through adding 2 μl of reduction buffer (500 mM TCEP (tis (2-carboxyethyl) phosphine) in 100 mM triethylammonium bicarbonate (TEAB)) for 1 h at 55 C and subsequently 2 μl of alkylation buffer (625 mM iodoacetamide in 100 mM TEAB) was applied for 1 h at room temperature in the dark. Protein concentration was measured by Direct Detect® Infrared Spectrometer (Merck, Germany) and proteins were subsequently digested with 60 ng μl −1 trypsin (Promega, V5111) at 37 C for 16 h. Tryptic peptides were dried in a vacuum concentrator and subsequently re-dissolved in aqueous 0.05% TFA/2% ACN. Three biological replicates were analysed based on two analytical replicates. As a control, we repeated the above-mentioned procedure without biotinylation of MH-S surface proteins to identify unspecifically bound proteins that bind to streptavidin beads.

LC-MS/MS analysis and protein database search
LC-MS/MS analysis was carried out on an Ultimate 3000 nano RSLC system coupled to a QExactive HF (I) and to a QExactive Plus (II) mass spectrometer (both Thermo Fisher Scientific) as previously described (Hoffmann et al., 2019). Peptides were trapped for 5 min on an Acclaim Pep Map 100 column (2 cm × 75 μm, 3 μm) at 5 μl min −1 followed by gradient elution separation on an Acclaim Pep Map RSLC column (50 cm × 75 μm, 2 μm). Eluent A (0.1% (v/v) formic acid in water) was mixed with eluent B (0.1% (v/v) formic acid in 90/10 acetonitrile/ water) as follows: (I): 0 min at 4% B, 6 min at 8% B, 30 min at 12% B, 75 min at 30% B, 85 min at 50% B, 90-95 min at 96% B, 95.1-120 min at 4% B. (II): 0-4 min at 4% B, 10 min at 7%, 40 min at 10% B, 80 min at 25% B, 90 min at 30% B, 110 min at 50% B, 115 min at 60% B, 120-125 min at 96% B, 125.1-150 min at 4% B. Positively charged ions were generated at 2.2 kV using a stainless steel emitter and a Nanospray Flex Ion Source (Thermo Fisher Scientific). The QExactive HF was operated in Full MS/data-dependent MS2 (Top15) mode. Precursor ions were monitored at m/z 300-1500 at a resolution of 120 000 (I)/70 000 (II) full-width at halfmaximum (FWHM) using a maximum injection time (ITmax) of 120 ms and an automatic gain control (AGC) target of 1e6. Precursor ions with a charge state of z = 2-5 were filtered at an isolation width of m/z 1.6 amu for HCD fragmentation at 30% normalized collision energy. MS2 ions were scanned at 15 000 (I)/17 500 (II) FWHM (ITmax = 90 ms (I)/120 ms (II), AGC = 2e5). The dynamic exclusion of precursor ions was set to 30 s. The LC-MS/MS instrument was controlled by Chromeleon 7.2 (I)/6.8 (II), QExactive Tune 2.8 and Xcalibur 4.0 (I)/3.0 (II) software. Tandem mass spectra were searched against the database of L. corymbifera 9682_v1 (21,812 sequence entries) (Schwartze et al., 2014b) as well as the UniProt database of Mus musculus (https://www.uniprot.org/proteomes/UP00000 0589, 2018/09/24; 54,109 sequence entries) by using Proteome Discoverer v1.4 and v2.2 and the search algorithms of Mascot 2.4 (Matrix Science, UK), Sequest HT (version of PD1.4 and PD2.2) and MS Amanda (1.0 and 2.0). PD1.4 was used for the biotinylated samples. Control samples without biotinylation were measured at a later time point with PD2.2. Two missed cleavages were allowed for the tryptic digestion. The precursor mass tolerance was set to 10 ppm and the fragment mass tolerance was set to 0.02 Da. Modifications were defined as static Cys carbamidomethylation and dynamic Met oxidation. Further dynamic modifications were set for carbamidomethylated thiopropanoyl groups (C 2 H 7 O 2 SN; +145 020 Da) derived from reductive cleavage and alkylation of Sulfo-NHS-SS-Biotin at biotinylated Lys residues. At least two peptides per protein and a strict false discovery rate (FDR) <1% were required for positive protein hits. The Percolator node and a reverse decoy database were used for q-value validation of spectral matches. Only rank 1 protein and peptides of the top-scored proteins were counted.

Gene ontology analyses of the protein populations identified by LC-MS/MS
Using the peptide list from the LC-MS/MS analysis, two quality criteria were applied to the PSMs value (i) For each biological replicate, peptides with PSM value above 15 and (ii) peptides with reported PSM value of 2 or higher in at least three technical replicates were considered. All keratins were removed from the candidate list because they are commonly known contaminants. Next, the overlaps of detected peptides across all conditions investigated were determined and visualized by a Venn diagram. Protein IDs were mapped to their associated genes using the R interface of BioMart (Durinck et al., 2005(Durinck et al., , 2009). The resulting gene lists were analysed using g: Profiler (Reimand et al., 2007) to obtain enriched GO terms. Redundant GO terms were filtered according to a previously described method (Kolte, 2019), in which Jaccard similarity coefficients were used to measure redundancy: Whereby A and B are significantly enriched gene sets. Using linear optimization, one gene set for each selected pair of highly overlapping gene sets (high Jaccard index) was discarded so that the number of genes represented by the remaining gene sets was maximal.
Gene-set enrichment analysis was performed with g: Profiler (Reimand et al., 2007) using gene set definitions from GO (Biological Process) (Ashburner et al., 2000). Significance levels were corrected for multiple testing using the Bonferroni method. For non-biotinylated macrophage proteins, the analysis was performed as described previously, but without removing keratins from the protein list. Next, the overlaps of detected peptides across all conditions investigated were determined and visualized by a Venn diagram.

Western blotting of biotinylated proteins
The identification of biotinylated proteins was carried out by Western blot as described previously (Schmidt et al., 2018). Briefly, 20 μg of biotinylated protein sample containing the denatured proteins that had bound to the spores of L. corymbifera was loaded on a 4%-12% (w/v) Bis-Tris Protein Gel (NuPAGE, Thermo Fisher Scientific) and separated by SDS-PAGE in 1× MES running buffer (Thermo Fisher Scientific) at 200 V for 40 min. The proteins were blotted with a semi-dry blot device (Bio-Rad) onto an Immobilon-FL PVDF membrane (Merck). The membrane was pre-hybridized with Western blocking reagent (Roche) for 1 h at room temperature and hybridized overnight with Alexa fluor® 488-conjugated IgG Fraction Anti-Biotin Antibody (Dianova, (1: 500; v/v) dilution in Tris-buffered saline (50 mM Tris, 150 mM NaCl, pH 7.5) for 90 min at room temperature and subsequently washed three times with transfer buffer solution (50 mM Tris-HCl, 150 mM NaCl, pH 7.5).
Fluorescence signals were detected by a Fusion FX7 system (Vilber Lourmat).

MALDI-TOF/TOF analysis
Excised protein bands were tryptically digested with 10 ng μl −1 of trypsin and generated peptides were subsequently extracted with 0.1% (w/v) TFA in acetonitrile as described previously (Shevchenko et al., 2007;Amarsaikhan et al., 2017). Peptides were spotted on an MPT 800/384 anchor chip target with saturated α-cyano-4-hydroxy-cinnamic acid (1:2 (v/v)), completely air-dried and measured by a MALDI-TOF/TOF ultrafleXtreme device (Bruker Daltonics, Bremen, Germany). MS and MS/MS spectra were searched against the NCBI database using the MASCOT 2.1.02 interface with the taxonomy setting 'Mus musculus' (Matrix Science, London, UK). The significance of hits was calculated based on the MASCOT score (p ≤ 0.05).

Detection of Hspa8 on the surface of macrophage cells and L. corymbifera spores by antibodies
In order to determine the presence of proteins on the surface of either macrophage cells or spores of L. corymbifera, we followed a previously described method of antibody-based detection (Liu et al., 2010). In brief, cell samples were blocked with 10% (v/v) normal goat serum with 1% (w/v) BSA for 1 h at room temperature. The cells were pelleted by centrifugation and washed three times with PBS-Ca 2+ /Mg 2+ . The cells were incubated with rabbit anti-Hspa8 antibody (Sigma-Aldrich, SAB2701964) (1:500; v/v) dilution in PBS for 90 min at room temperature or overnight at 4 C. The cells were stained with goat anti-rabbit IgG Alexa Fluor 647 antibodies (5 μg/ml, Thermofisher, A-21245) for flow cytometry or with goat antirabbit IgG H&L Alexa Fluor 488 (Abcam, ab150077) for confocal laser scanning microscopy analysis for 90 min at room temperature. In order to determine a possible colocalization of Hspa8 with phagocytosed L. corymbifera spores, the macrophage cells were permeabilized with 1% (v/v) Triton X-100 for 10 min at room temperature, stained with anti-Hspa8 antibody, and subsequently examined using immunofluorescence.

Measurement of phagocytosis
Monitoring of the phagocytosis of fungal spores by macrophage cells was carried out as previously described (Kraibooj et al., 2014) with some modifications. Spores were stained with FITC (Sigma-Aldrich Chemie GmbH, Germany) in 0.1 M Na 2 CO 3 at 30 C for 30 min, centrifuged at 5000g for 5 min and then washed three times with PBS to remove excessive FITC. Then, the number of spores was determined by manual counting with a Neubauer chamber. Finally, spores were diluted to a final concentration of 10 6 in 500 μl of RPMI medium.
2 × 10 5 macrophage cells were cultivated on glass coverslips (12 mm Ø) in 24-well plates (NUNC) in RPMI-1640 and were incubated in a CO 2 incubator (5% (v/v) CO 2 ) overnight at 37 C to ensure adherence of the macrophage cells to the coverslips. The FITCstained spores were confronted with macrophage cells at an MOI of 5 and centrifuged with 100g for 5 min to start the interaction between spores and macrophage. After 1 h, the interaction was stopped by removing the non-adherent spores from the supernatant by washing three times with ice-cold PBS. About 500 μl of calcofluor white (CFW; Sigma-Aldrich) was added for 10 min at room temperature to stain spores, which adhered to the cell membrane of macrophage cells. The cells were washed two times with PBS to remove excessive CFW. The cells were fixed by the addition of 500 μl 3.7% (v/v) formaldehyde in PBS for 15 min at room temperature.
Roti-mount FluorCare (ROTH, HP19.1) mounting medium was used to increase the resolution and to prohibit instant bleaching of the samples during microscopy. Microscopic images were taken by the Axio Observer 7 Spinning Disk Confocal Microscope (ZEISS, Jena, Germany) and processed with ZEN 2.1 Software (ZEISS). At least seven tile scans (3 × 3) images were documented for each sample. For the blocking assay, each well of macrophage S cells was incubated with 50 μg of rabbit Anti-Hspa8 antibody (Sigma-Aldrich, SAB2701964) or with rabbit IgG Isotype Control (Thermo Fisher, 02-6102) before confrontation with spores. Microscopic analysis was performed from three biological replicates for each strain.

Automated image analysis of immunofluorescence images
The images were analysed using a custom-designed toolkit written in the macro language of ImageJ using the Fiji platform (Schindelin et al., 2012). The automated workflow was based on objective and quantitative systems biology methods (Figge and Murphy, 2015;Medyukhina et al., 2015;Figge, 2018) applied specifically to host-pathogen interactions (Mech et al., 2011(Mech et al., , 2014Mattern et al., 2015;Cseresnyes et al., 2018). The toolkit was able to segment the phagocytes purely based on their transmitted light images, thus avoiding the costly and biologically interfering fluorescence-labelling process. The macrophage cells were identified by enhancing the contrast of the image pixels that were part of the cells' edges. This was achieved by applying a two-dimensional Hessian filter to the transmitted light bright field images in order to identify the outline of the macrophage cells. The Hessian filtering process resulted in two Eigen matrices that were visualized as images. Out of the two Hessian Eigen matrices, the smallest eigenvalues were used to detect the edges of the macrophage cells that were calculated for each Z layer of the 3D images. The resulting Hessian Z stack of images was utilized to reconstruct the macrophage cells as 3D surfaces. The phagocytosed and adherent spores were identified via fluorescence labelling, whereas the images of the green fluorescence channel determined the location of the FITC labelled spores. The blue channel identified the adherent spores via CFW staining. By calculating the number of green and blue fungal spores that overlapped with the Hessiansegmented macrophages, it was possible to calculate the main phagocytic measures, as described in Kraibooj et al. (2015), Cseresnyes et al. (2018) and .

Measuring the intensity of Hspa8 on the surface of macrophages after applying different MOIs
To measure whether the distribution of Hspa8 on the surface of macrophage cells was influenced by the MOI used, we followed the methods described in a previous study with minor modification (Ziegler et al., 2017). Briefly, the phagocytosis assay was performed as explained above (Measurement of phagocytosis) but using different MOIs (1, 3 and 5) for flow cytometry analyses. After fixation of the cells with 3.7% (v/v) formaldehyde, staining was carried out with anti-Hspa8 antibody (1:500; v/v) dilution in PBS for 90 min at room temperature, and subsequently incubated with goat anti-rabbit IgG Alexa Fluor 647 secondary antibody (5 μg ml −1 ; w/v) in PBS-Ca 2+ /Mg 2+ for 90 min at room temperature. The intensity of the Hspa8 positive fluorescent signal was determined by the measurement of the emission intensity of the Alexa Fluor 647 fluorophore.

Statistical analysis
The non-parametric Wilcoxon rank-sum test was used to calculate the significance of Hspa8 on the surface of macrophage cells in the recognition of L. corymbifera spores, the blocking assay, and the effect of MOI on the expression of Hspa8 on the surface of macrophage cells. p-values less than 0.05 were considered significant and marked as *, ** indicates p < 0.01 and ***p < 0.001 (Wilcoxon, 1945).