Biotinylated Surfome Profiling Identifies Potential Biomarkers for Diagnosis and Therapy of Aspergillus fumigatus Infection

Aspergillus fumigatus is the most important airborne human-pathogenic mold, capable of causing both life-threatening invasive pulmonary aspergillosis in immunocompromised patients and allergy-inducing infections in individuals with atopic allergy. Despite its obvious medical relevance, timely diagnosis and efficient antifungal treatment of A. fumigatus infection remain major challenges. Proteins on the surface of conidia (asexually produced spores) and mycelium directly mediate host-pathogen interaction and also may serve as targets for diagnosis and immunotherapy. However, the similarity of protein sequences between A. fumigatus and other organisms, sometimes even including the human host, makes selection of targets for immunological-based studies difficult. Here, using surface protein biotinylation coupled with LC-MS/MS analysis, we identified hundreds of A. fumigatus surface proteins with exposed regions, further defining putative targets for possible diagnostic and immunotherapeutic design.

All the aforementioned methods efficiently extract surface proteins and/or peptides, but they also come with drawbacks; they partially disrupt the surface layer and potentially release cell wall-embedded and cytosolic proteins and peptides in addition to their target cell surface proteins. Cell surface biotinylation has been shown to yield a low rate of contamination by cytoplasmic proteins (24) and in particular provides useful information about the surface-exposed protein regions.
Here, we used a widely applied amine-reactive biotinylation method for surface proteomics (16,25,26) to label the exposed lysine or N-terminal amino acid residues (AAs) of A. fumigatus cell surface proteins. Subsequently, liquid chromatographytandem mass spectrometry (LC-MS/MS) analysis was applied to detect streptavidinenriched proteins and to define surface-exposed regions based on lysine modifications. The surface localization of several detected proteins was verified through immunofluorescence of Myc-tagged transformants of A. fumigatus. Our report complements the picture of the A. fumigatus surface proteome and reveals surface proteins that may serve as detection markers or targets for immunotherapy in the future. Most importantly, we may have identified a number of proteins that are likely directly involved in the interaction of A. fumigatus and the human host.

RESULTS
Biotinylation of A. fumigatus surface proteins. In this study, we sought to expand the repertoire of A. fumigatus surface proteins and their surface-exposed epitopes, which could directly mediate pathogen-host interaction and serve as targets for diagnosis and therapy. We used the biotinylation reagent sulfosuccinimidyl-6-(biotinamido)hexanoate (Sulfo-NHS-LC-Biotin) to label the surface-exposed primary amine residues (e.g., the -amino group of lysine residues) throughout germination of A. fumigatus conidia (Fig. 1). The water-soluble and negatively charged reagent Sulfo-NHS-LC-Biotin is one of the most frequently applied biotinylation reagents, particularly in regard to surface proteomics (25)(26)(27). Samples were collected from dormant conidia grown at 37°C on Aspergillus minimal medium (AMM) agar plates, including swollen conidia (5 h), germlings (8 h), and hyphae (14 h) germinated at 37°C in RPMI liquid medium, as described previously (11). Although there is a possibility of biotin reagent permeation into the fungal cells (26), the biotinylation signal was confined mainly to the surface of conidia and hyphae as observed by immunofluorescence staining against biotinylated proteins ( Fig. 2A; see also Fig. S1 in the supplemental material). Due to the high rigidity of the A. fumigatus cell wall, the fungal . Some noncovalently linked cell wall proteins were extracted using SDS buffer. Biotinylated proteins were purified using magnetic streptavidin beads and then analyzed by LC-MS/MS. samples were first disrupted by glass beads in phosphate-buffered saline (PBS) buffer to release loosely attached surface proteins. This was followed by a second step using SDS buffer extraction to remove non-covalently bound hydrophobic surface proteins (Fig. 1). After purification, the samples were checked for the level of protein biotinylation and recovery from the streptavidin beads by Western blotting (Fig. 2B; see also Fig. S2). Proteins that differed widely in molecular mass were detected, demonstrating the ability to label a wide range of proteins using this approach. We also observed biotinylated protein bands in nonbiotinylated control samples ( Fig. 2B; see also Fig. S2). Most likely, these represented the three known biotin-dependent carboxylases in A. fumigatus, encoded by Afu4g07710, Afu2g08670, and Afu5g08910, each of which showed high abundance according to the observed peptide spectrum matches (PSM; mass spectral count of repeated peptide identifications that serve as an approximate measure of the relative protein abundance) in our LC-MS/MS analyses (see Data Set S1 in the supplemental material). They are easily distinguishable from the NHS-LC-Biotin modification, which adds 339 Da of mass to the modified proteins. a detectable NHS-LC-Biotin modification (Fig. 3A). Note that many of these proteins were also detected in the nonbiotinylated control samples without biotinylation modifications (see Data Set S1). Fifty-nine (26%) biotinylated proteins had a predicted signal peptide, and 16 (7%) proteins had at least one transmembrane domain (see Table S1 in the supplemental material) (https://fungidb.org/fungidb/ [29]), which demonstrates the enrichment of extracellular proteins in the biotinylated fraction.
Over the course of germination, 61 proteins with biotinylation were identified on dormant conidia, 69 on swollen conidia, 85 on germlings, and 136 on hyphae. Most of the proteins detected with biotin modifications were found in PBS extracts; only a few (5 to 14) were exclusively found in SDS extracts ( Fig. 3A and B; see also Data Set S1). In accordance with our previous surface proteomics study based on a trypsin-shaving approach (11), our data confirm the dynamic change of the surface-exposed proteome of A. fumigatus across germination. There were 16, 17, 33, and 82 proteins detected exclusively on dormant conidia, swollen conidia, germinating conidia, and hyphae, respectively (Fig. 3C). The very different surfome compositions that we observed compared to our previous study (11) can potentially be explained in several ways. First, trypsin is more prone to disturbing the fungal surface structure and thereby releasing peptides of proteins embedded in the cell wall; second, some surface proteins may not be accessible to enzymatic cleavage or biotinylation reactions; third, proteins are inefficiently released from streptavidin beads.
Htb1, Tef1, and Afu5g10550 were also abundantly detected throughout germination ( Table 2). Somewhat surprisingly, the histone H2B (Htb1, Afu3g05350) seemed to be one of the most abundant proteins throughout the germination time course (Table 1 and 2). The number of PSM/length of Htb1 in dormant conidia (3.25) was even higher than that determined for RodA (2.37) ( Table 2), in contrast to previous studies performed using different approaches (10,11,21). One possible explanation for this discrepancy is the difference in the total number of the modifiable lysines in each protein. There are 23 lysine residues in the 140 amino acid residues (AAs) of Htb1, 9 of which were detected with a biotinylation modification (Table 1). In contrast, there are only 9 lysine residues present in the hydrophobic protein RodA. RodA has been shown to be present throughout germination by previous studies (11,21); this could be explained by contamination with ungerminated conidia or by a low level of RodA in hyphae. In this study, RodA was abundantly detected in dormant, swollen, and germinating conidia ( Table 2). Three sites, including K50, K55, and K126, were detected with biotinylation marks. All three lysine residues are localized in the relatively hydrophilic ␣-helical regions of the protein (33), suggesting that only the surface-exposed lysine residues are likely to be biotinylated.
Immunoreactive proteins are exposed on the surface of A. fumigatus. In a unique subset of individuals with atopic allergy, sensitization to A. fumigatus allergens can develop into allergic asthma, allergic sinusitis, and, following fungal lung colonization, into ABPA (1). Twenty-three different allergens have been reported in accordance with the systematic allergen nomenclature (www.allergome.org), but actually more than 100 immunoreactive A. fumigatus proteins have already been uncovered by immunoproteomic studies (5). In our work, 20 allergens (  (Table 3). Consistent with the literature, these data again clearly demonstrate that numerous known A. fumigatus allergens are exposed on the fungal surface (9-11, 19, 34).
In addition to the allergens, there were also other immunoreactive proteins that could serve as biomarkers for diagnosis or targets for immunotherapy (35)(36)(37). Our analysis revealed biotinylation of DppV, Bgt2, and Afu5g10550 in all four morphotypes (Table 1). Afu5g10550 encodes an ATP synthase F1 beta subunit, which reacts with immunosera from rabbits exposed to A. fumigatus conidia (35). Biotinylation of Afu5g10550 K122 was detected throughout germination. It was also one of the most Histone H2B 140 Putative dihydrolipoamide dehydrogenase 513 Has domain(s) with predicted DNA binding, protein heterodimerization activity 265 Putative triosephosphate isomerase 249 a Lysine residues of the protein detected with biotinylation in dormant (D), swollen (S), or germinating (G) conidia or hyphae (M) and not detected (-) are indicated.
prevalent proteins found on dormant, swollen, and germinating conidia (Table 1 and 2). Immunoreactive GpdA, Asp f MDH, Bgt1, Asp f 6, Hsp70, PgkA, Asp f 4, GliT, and Asp f GT proteins were also prevalent on the surface of one or two morphological stages ( Table 2). Heat shock proteins are exposed to the surface. Heat shock proteins (HSPs) and a large set of cochaperones are ubiquitous molecular chaperones that act in maintaining protein homeostasis (38). Although they represent highly abundant housekeeping proteins, it has been well established that HSPs are also secreted extracellularly or localized on the cell surface (39), where they exhibit moonlighting function. Hsp90, also known as Asp f 12, is detectable on the cell wall of A. fumigatus (34) and was also found to be biotinylated in this study (Table 3). Considering the detection of biotinylated Hsp90 and Hsp70, we investigated whether other chaperone-related proteins are present on the cell surface of A. fumigatus. Indeed, at least 14 chaperone-related proteins were detected with biotinylation modifications ( Table 4). Most of the chaper- ones were detected on the surface of germlings or hyphae with the exception of Scf1 and GrpE. Scf1 shows similarities to the 12-kDa heat shock protein of Saccharomyces cerevisiae and was biotinylated at K81 (Table 4). It was found to be one of the most prevalent proteins on dormant conidia of A. fumigatus strains CEA10 (Table 2) and ATCC 46645 (10). Five chaperones (Hsp70, Hsp88, HscA, BipA, and Ssc70) belonging to the Hsp70 family were detected in our study (Table 4). In addition to Hsp70, four other HSPs (Sti1, Hsp88, HscA, and Ssc70) have been shown to induce serological antibody responses in ABPA patients (37).
To confirm the surface localization of the HSPs, we generated constructs by fusing a Myc tag to the C terminus of the HSPs. These constructs were ectopically integrated into the genome of A. fumigatus (Table S2 and S3). The hsp70-Myc, ssc70-Myc, bipA-Myc, and ssz-Myc transformants were verified by immunoblotting using an anti-Myc tag antibody (Fig. S4). Using immunofluorescence microscopy, we found that all of the  Myc-tagged fusion proteins were localized on the surface of germinating conidia (Fig. 4A) and hyphae (Fig. 4B), except for the negative control, the HSP70 chaperone Ssz (Afu2g02320), which was not detected on the surface in our biotinylation experiment. This independently confirms that three HSP70 chaperones of A. fumigatus were able to be localized to the cell surface of the fungus. The core surface proteome (surfome) of A. fumigatus. Considering that several surface proteomic studies have been performed using different methods (trypsin shaving, HF/pyridine, and formic acid extraction), we attempted to compare these data sets to provide a more comprehensive picture of the A. fumigatus surfome (9-11). In total, 946 different proteins were detected as surface proteins in this study and the aforementioned studies, including 416 proteins that were detected by at least two approaches ( Fig. 5; see also Table S1). A total of 39 proteins were commonly identified in all studies (Fig. 5) (Table 5; see also Table S1). These 39 proteins included 15 of the most prevalent proteins identified in our study and 9 allergens (Tables 2, 3, and 5;  see also Table S1). Two small proteins, Grg1 (glucose-repressible gene) and ConJ (conidiation-specific protein 10), were detected among the most prevalent proteins on dormant conidia of the ATCC 46645 strain using the HF-pyridine extraction method (10), on the CEA10 strain using the trypsin-shaving method (11), and in this study using biotinylation (Table 2 and 5). Grg1 was detected with biotinylation at the positions K32 and K46, while ConJ was biotinylated at the N terminus and site K48 ( Table 2 and 5). Conidial surface protein CcpA, which contributes to fungal virulence, was detected with biotinylations at positions K41 and K90 (Table 5). Putative glycosylphosphatidylinositol (GPI)-anchored cell wall protein CweA (11) was detected with a biotinylation marker at site K358 (Table 5). In addition to the 39 commonly detected proteins, 67 proteins were detected in at least four surfome data sets (Fig. 5A), 39 of which were detected with biotinylated amino groups (Table S1) Table 2). In total, 183 surface proteins with biotinylation modifications identified in this study were also detected by alternative methods (Fig. 5B). All in all, the core surfome of A. fumigatus provides a valuable database of protein targets for further studies on host-pathogen interactions, diagnosis, and immunotherapy.

DISCUSSION
Proteins in combination with other cell wall components on the surface of A. fumigatus conidia play a key role in protecting the fungus from environmental insults and host defense responses during an infection (10,12,16,40,41). Several methods and techniques have been used to investigate the surface proteins of this pathogenic fungus, leading to the detection of hundreds of surface proteins, whose presence on the surface changes dynamically during development (9-11, 20, 23). Most methods used for the extraction of surface proteins, such as enzymatic treatment with glucanase or trypsin or acidic extraction with formic acid or hydrogen fluoride-pyridine, have the potential to release intracellular or unexposed cell wall proteins. Cell-impermeable biotinylation reagents, which react with primary amines accessible on the cell surface, have successfully been used for labeling and identification of surface proteins with little contamination (16,25,26,42,43). To further expand the range of A. fumigatus surface proteomes and to solidify the data representing the common core surfome of A. fumigatus, we used a surface biotinylation approach, which had not been applied to A. fumigatus previously. The aim was to characterize and detail the A. fumigatus surfome with surface-exposed regions across germination and verify the surface localization of selected proteins by an additional method (Fig. 4). Therefore, our work provides a multitude of candidates for further investigation of host-pathogen interaction and possible immunodiagnostic/therapeutic usages.
Proteins on the surface mediate direct contacts between pathogens and hosts. In addition to several known surface proteins, such as RodA, CalA, Asp f 2, and CcpA (10,12,16,19), abundant surface proteins detected in this study, such as Tef1, ArtA, Hsp70, and Hsp90, also have the potential to interact with a range of host proteins (44). Candida albicans Tef1 was shown to be surface localized and to be able to bind human  Table 5), including the results obtained in this study (CEA10 B, corresponding to all the morphotypes of CEA10 obtained using the biotinylation method); the surfome of CEA10 (CEA10 T, corresponding to all the morphotypes of CEA10 obtained using the trypsin-shaving method); D141 (dormant and swollen conidia) results obtained using the trypsin shaving method (10,11); the CEA17ΔakuB KU80 dormant conidia surfome results obtained using the formic acid extract (9); and the ATCC 46645 surfome detected by hydrogen fluoride-pyridine extraction and trypsin shaving (10). The pink zones highlight the number of proteins present in at least two sets of surfome data but not all sets of surfome data. Gray zones indicate the number of proteins present in just one experiment.  Tables 1 and 2 ArtA  Table 1 Bgt2  Table 1 Asp f MDH   Table 2 GpdA   Table 3 Asp f gamma_actin (Act1)

Afu6g04740
Actin 393 K346 M See Table 3 Asp f mannosidase (MsdS)  Table 3 Asp f catalase (Cat1)  Table 3 Asp f glucosidase (Exg12)  Table 3 Nhp6 plasminogen, probably through C-terminal lysine residues (45). In this study, Tef1 was detected with biotinylation at the C-terminal K472-to-K486 region throughout germination (Table 1), indicating that Tef1 in A. fumigatus might have a similar function. The 14-3-3 proteins, such as ArtA, also have the ability to bind a multitude of proteins and to play an important role in morphogenesis and sensing of environmental cues in fungi (46)(47)(48). The presence of HSPs on the surface of mammalian cells and microorganisms has been well documented (39,(49)(50)(51)(52)(53) and was also confirmed in this study (Fig. 4). The surface-localized HSPs influence the interactions between viral, bacterial, and fungal pathogens and with host cells as well (49,(54)(55)(56). For example, expression of Listeria monocytogenes heat shock protein ClpC, a member of the 100-kDa heat shock protein family, is required for cell adhesion and invasion (53) and also allows this bacterium to escape from the phagosome (52). C. albicans Hsp70 protein Ssa1 is an invasin that binds to host cell cadherins to induce host cell endocytosis, which is critical for C. albicans to cause maximal host cell damage (51). In line with this, A. fumigatus Hsp70 and Hsp90 were predicted to interact with host proteins in conidia containing mouse macrophage phagolysosomes (44), suggesting the potential roles of surface HSPs in manipulating the host immune responses.
In addition to their roles in pathogenicity, surface proteins could also be targets of immunotherapies based on the use of antibodies. Surface proteins are particularly suitable for such use due to their accessibility. The use of antibody-based therapies is rapidly growing, since they represent a promising approach for directly attacking the pathogen and boosting the innate immune system. However, only a few monoclonal antibodies against fungal pathogens have been developed and have advanced to clinical trials (57). One example is Mycograb, an Hsp90-specific antibody fragment, which showed promise for treating Candida infections in combination with amphotericin B (58) but failed to obtain marketing authorization. In mouse experiments, treatment with anti-Hsp60 antibodies reduced fungal burden after infection with the dimorphic fungi Histoplasma capsulatum and Paracoccidioides lutzii (59,60). On the negative side, the amino acid sequences of HSPs are highly conserved and crossreactivity can occur. For example, the epitope (NKILKVIRKNIVKK) that Mycograb targets shows high similarity between yeast, mice, and human homologues of Hsp90 (58), including A. fumigatus Hsp90 (NKIMKVIKKNIVKK, amino acids 383 to 396). Other abundant, fungus-specific surface proteins may represent better targets for immunotherapy as discussed in the following section.
To date, more than 100 antigens or immune-reactive proteins of A. fumigatus that react with sera from ABPA patients or animal models have been identified using classical immunobiological procedures (35)(36)(37). However, only a few recombinant allergens have been used commercially for diagnosis of allergic aspergillosis (5). Although the crystal structures of some allergens are known (61,62), the issue of whether the allergen/protein exhibits special structural characteristics that are responsible for its allergenicity is still poorly understood. Thus, information is needed about the association of allergens with the different morphotypes (conidia, mycelium) and their exposed regions that may directly mediate the interaction with host components, such as IgE binding. Such knowledge creates the basis for understanding the immunological properties of protein antigens and is important for the establishment of new forms of diagnosis and treatment (63).
The cyclophilins, including Asp f 11 and Asp f 27 (62), are structurally conserved pan-allergens able to elicit IgE-mediated hypersensitivity reactions (62,64,65). It was reported previously that the conserved N81-N149 region of Rhizopus oryzae cyclophilin Rhi o 2 (see Fig. S5 in the supplemental material) is crucial for IgE recognition and cross-reactivity (64). In this study, however, the C-terminal region of Asp f 27, which is not conserved, was detected with biotinylation marks instead (Table 3; see also Fig. S5). This also provides a possible target for immunotherapy.
In addition to the cyclophilin allergens, several other allergens were detected with biotinylation sites on different morphotypes, such as the K50 -K101 region of Asp f 6, the K304 -K327 region of Asp f FDH, K238 -K269, and the K303-K331 region of Asp f MDH ( Table 3). Note that there were also regions of surface proteins that were detected on all morphotypes. For example, the K472-K486 region of Tef1, the K122 region of Afu5g10550, the K225-K231 region of Bgt2, and the K420 -K435 region of the putative dihydrolipoamide dehydrogenase Afu2g02100 were all consistently surface exposed (Table 1). These peptides may represent promising candidate antigens for the development of monoclonal antibodies to be used for diagnosis and immunotherapy.
As published in our previous studies, the dynamic surfome of A. fumigatus is affected by various cultivation conditions, including the medium, temperature, and time, and contains many proteins secreted by nonclassical pathways (10,11). Correspondingly, a recent study also revealed that the phenotypes of A. fumigatus germinating conidia vary among genetically identical conidia (66). Here, we detected 61, 69, and 85 proteins on dormant, swollen, and germinating conidia, respectively; while 136 proteins were found on the surface of hyphae (Fig. 3C). The comparison of the different surface proteomics data sets helped to identify the core surfome of A. fumigatus (Fig. 5). These proteins, which are consistently found on the surface of A. fumigatus, likely play a role in mediating the interaction of A. fumigatus with the host or other organisms.

MATERIALS AND METHODS
Fungal strains and cultivation. All strains used in this study are listed in Table S2 in the supplemental material. A. fumigatus CEA10 was cultivated as described previously (11). Briefly, the CEA10 strain was inoculated on Aspergillus minimal medium agar plates with 1% (wt/vol) glucose for 7 days at 37°C. Conidia were harvested in sterile H 2 O and separated from hyphae and conidiophores by filtering (30-m pore size; Miltenyi Biotec). For germination of swollen conidia, 1 ϫ 10 10 freshly collected conidia were incubated shaking in 50 ml RPMI 1640 (Lonza) for 5 h at 37°C. A total of 1 ϫ 10 9 conidia were germinated under the same set of conditions for 8 h to enrich for germlings and 1 ϫ 10 9 conidia in 100 ml RPMI 1640 for 14 h to enrich for hyphae.  (25,26). We chose a shorter time, since a longer incubation time or an incubation at elevated temperatures could increase the leakage of the biotinylation reagents inside the cell. Experiments were performed as described previously (26) with minor modification. Conidia, germlings, and hyphae were washed three times with PBS (pH 7.4) and then incubated in 5 ml of PBS containing 5 mg EZ-Link Sulfo-NHS-LC-Biotin (Thermo Fisher Scientific, catalog no. f21335) for 1 h at 4°C. As a nonbiotinylated control, fungal materials were also incubated in PBS alone. In all cases, the reaction was terminated by adding two volumes of 100 mM Tris-HCl (pH 7.4), and the reaction mixture was incubated for a further 30 min. The samples were then washed another three times with PBS (pH 7.4). All experiments were performed in three (dormant conidia, swollen conidia) or four biological (germlings, mycelia) replicates.
Protein extraction and purification. After addition of 1 ml of PBS (pH 7.4) containing protease inhibitor (Roche cOmplete; catalog no. f04693159001) (1 tablet per 10 ml) and 500 l of 0.5-mm-diameter glass beads, conidia, germlings, and hyphae were disrupted using a FastPrep homogenizer with the following settings: 6.5 m/s, 3 times for 30 s each time. The samples were then centrifuged at 16,000 ϫ g for 10 min at 4°C. The supernatants were collected and denoted "PBS extracts." The pellets were washed twice with 1 M NaCl and another three times with 50 mM Tris-HCl (pH 7.4) followed by extraction with SDS buffer ( Western blot analysis. Proteins were separated by SDS-PAGE using NuPAGE 4%-to-12% (wt/vol) Bis-Tris gradient gels (Thermo Fisher Scientific) and were then transferred onto a polyvinylidene difluoride (PVDF) membrane using an iBlot 2 dry blotting system (Thermo Fisher Scientific). To detect biotinylated proteins, 5 g of total protein and 20 l of purified protein were loaded and membranes were blocked with Western blocking reagent (Roche) and then incubated with Pierce streptavidinhorseradish peroxidase (HRP) (Thermo Fisher Scientific, catalog no. f21130) (1:2,000) overnight at 4°C. To detect the Myc-tagged proteins, a 10-g volume of total protein was loaded and membranes were blocked with a 5% (wt/vol) solution of skim milk and then incubated with primary antibody (Myc-tagged mouse monoclonal antibody [MAb]; Cell Signaling Technology, catalog no. f2276) (1:1,000) overnight at 4°C. Hybridization with a secondary antibody (HRP-linked anti-mouse IgG; Cell Signaling Technology, catalog no. f7076) was performed for 1 h at room temperature. Chemiluminescence of HRP substrate was detected with a Fusion FX7 system (Vilber Lourmat, Germany).
Immunofluorescence microscopy. The fungal materials were washed three times with PBS (pH 7.4) and then blocked with 1% (wt/vol) bovine serum albumin (BSA)-PBS for 1 h at room temperature. For detection of surface biotinylation, A. fumigatus conidia, germlings, and hyphae were incubated with Alexa Fluor 635-conjugated streptavidin (Thermo Fisher Scientific, catalog no. S32364) (1:100) for 1 h in the dark. The surface localization of Myc-tagged proteins was examined using an anti-Myc primary antibody (Myc-tagged mouse MAb; Cell Signaling Technology, catalog no. f2276) (1:100) for 2 h at room temperature and a secondary antibody (donkey anti-mouse IgG H&L Alexa Fluor 568, Abcam catalog no. fab175472) (1:500) for 1 h at room temperature in the dark. After three washes with PBS, the samples were examined under a Zeiss Axio Imager M2 microscope.
In-solution protein digestion with trypsin. Eluted protein samples (150 l) were reduced by adding 4 l of 500 mM TCEP [Tris(2-carboxyethyl)phosphine]-100 mM TEAB (triethylammonium bicarbonate) for 1 h at 55°C. A 4-l volume of 625 mM iodoacetamide was added to each sample followed by incubation for 30 min at room temperature in the dark. Proteins were then precipitated using the methanol-chloroform-water method (67). Protein samples were resolubilized in 100 l of 100 mM TEAB and sonicated for 10 min. The protein content was measured with a Merck Millipore Direct Detect infrared spectrometer. The samples were treated with trypsin/Lys-C protease mix (Promega, catalog no. fV5072) at a protease/protein ratio of 1:25 for 16 h at 37°C. The reaction was stopped with 10 l of 10% (vol/vol) formic acid, and the reaction mixture was subjected to evaporation using a SpeedVac (Thermo Fisher Scientific). Peptides were resolubilized in 25 l 0.05% (vol/vol) trifluoroacetic acid (TFA)-2% (vol/vol) acetonitrile (ACN)-water and sonicated for 15 min in a water bath before transfer to a 10-kDa-molecular-weight cutoff filter. After 15 min of centrifugation at 14,000 ϫ g (4°C), the samples were transferred to high-performance liquid chromatography (HPLC) vials and stored at -20°C.
Genetic manipulation of A. fumigatus. All oligonucleotides utilized in this study are listed in Table S3. For surface localization studies, a pLJ-Ssc70-Myc construct was generated using plasmid pTH1 (68) as the backbone. A 3,276-bp DNA fragment containing a 1,010-bp 5= promoter region, the ssc70 gene without TAA, and an in-frame fused Myc-tagged coding sequence was PCR amplified from genomic DNA using primers Ssc70-M_F and Ssc70-M_R. Another 371-bp DNA fragment was amplified from plasmid pTH1 using primers Myc_F and pTH_R. These two fragments were mixed as the template and amplified using primers Ssc70-M_F and pTH_R to generate the Ssc70-Myc fragment. Ssc70-Myc fragment was digested with KpnI and NotI and then inserted into the pTH1 plasmid using T4 ligase (Thermo Fisher Scientific) to yield the pLJ-Ssc70-Myc plasmid. The pLJ-Ssc70-Myc plasmid was then used as the backbone to generate other Myc-tagged constructs. For example, a 3,073-bp DNA fragment containing the hsp70 gene and the promoter region was PCR amplified from genomic DNA using primers Hsp70-M_F and Hsp70-M_R. The DNA fragment was then inserted into the KpnI-and HindIII-digested pLJ-Ssc70-Myc backbone using a CloneEZ PCR cloning kit (GenScript) to get plasmid pLJ-Hsp70-Myc. Primers BipA-M_F and BipA-M_R were used to amplify the bipA gene with a promoter region. Primers Ssz-M_F and Ssz-M_R were used to amplify the ssz gene with a promoter region. The plasmids were used to transform protoplasts from A. fumigatus strain A1160 (CEA17 ΔakuB KU80 ) (69).
LC-MS/MS analysis. LC-MS/MS analysis of tryptic peptides was performed on an Ultimate 3000 RSLC nano instrument coupled to a Q Exactive HF mass spectrometer (both Thermo Fisher Scientific) in two analytical replicates. Tryptic peptides were trapped for 4 min on an Acclaim Pep Map 100 column (2 cm by 75 m, 3-m pore size) at a flow rate of 5 l/min. The peptides were then separated on an Acclaim Pep Map column (50 cm by 75 m, 2-m pore size) using a binary gradient (A, 0.1% [vol/vol] formic acid-H 2 O; B, 0.1% [vol/vol] formic acid-90:10 [vol/vol] ACN/H 2 O) as follows: 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 to 95 min at 96% B, and 95.1 to 120 min at 4% B. Positively charged ions were generated by the use of a Nanospray Flex ion source (Thermo Fisher Scientific) using a stainless steel emitter with 2.2-kV spray voltage. Ions were measured in datadependent MS2 (Top15) mode. Precursor ions (z ϭ 2 to 5) were scanned at m/z 300 to 1,500 (resolution [R], 120,000 full width at half maximum [FWHM]; automatic gain control [AGC] target, 3·10 6 ; maximum injection time [IT max ], 120 ms). Fragment ions generated in the higher-energy collisional dissociation (HCD) cell at 30% normalized collision energy using N 2 were scanned (R, 15,000 FWHM; AGC target, 2·10 5 ; IT max , 90 ms) using a dynamic exclusion duration of 30 s.
Database search and data analysis of trypsin-cleaved surface peptides. The MS/MS data were searched against the Aspergillus Genome database (AspGD) entries for Aspergillus fumigatus Af293 (http://www.aspergillusgenome.org/download/sequence/A_fumigatus_Af293/current/; 3 February 2019) using Proteome Discoverer (PD) 2.2 and the algorithms of Mascot 2.4.1, Sequest HT, and MS Amanda 2.0. Two missed cleavages were allowed for tryptic peptides. The precursor mass tolerance was set to 10 ppm, and the fragment mass tolerance was set to 0.02 Da. Dynamic modifications were set as oxidation (ϩ15.995 Da) of Met and NHS-LC-Biotin (ϩ339.162 Da) modification of Lys and the protein N terminus. The static modification was set to carbamidomethylation (ϩ57.021 Da) of Cys. One unique rank 1 peptide with a strict target false-discovery (FDR) rate of Ͻ1% on both the peptide and protein levels (compared against a reverse-decoy database) was required for positive protein hits. Only proteins identified in at least two different biological replicates were considered. Protein abundance quantification was performed by the use of the Minora algorithm of PD2.2 (area under the curve approach).
Data availability. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE (70) partner repository with the data set identifier PXD018071.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.