Proteome Analysis Reveals the Conidial Surface Protein CcpA Essential for Virulence of the Pathogenic Fungus Aspergillus fumigatus

The mammalian immune system relies on recognition of pathogen surface antigens for targeting and clearance. In the absence of immune evasion strategies, pathogen clearance is rapid. In the case of Aspergillus fumigatus, the successful fungus must avoid phagocytosis in the lung to establish invasive infection. In healthy individuals, fungal spores are cleared by immune cells; however, in immunocompromised patients, clearance mechanisms are impaired. Here, using proteome analyses, we identified CcpA as an important fungal spore protein involved in pathogenesis. A. fumigatus lacking CcpA was more susceptible to immune recognition and prompt eradication and, consequently, exhibited drastically attenuated virulence. In infection studies, CcpA was required for virulence in infected immunocompromised mice, suggesting that it could be used as a possible immunotherapeutic or diagnostic target in the future. In summary, our report adds a protein to the list of those known to be critical to the complex fungal spore surface environment and, more importantly, identifies a protein important for conidial immunogenicity during infection.

as a structural protein impacting conidial immunogenicity rather than possessing a protein-intrinsic immunosuppressive effect. Together, these data suggest that CcpA serves as a conidial stealth protein by altering the conidial surface structure to minimize innate immune recognition.
conidial cell-surface protein required for virulence, possibly via alteration of conidial innate immune recognition.

RESULTS
Conidial surface proteome analysis reveals the highly abundant CcpA protein. To comprehensively elucidate the conidial surface proteome of A. fumigatus, we performed liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis of conidial cell wall proteins extracted by hydrogen-fluoride (HF)-pyridine treatment as well as of surface-exposed conidial peptides cleaved by trypsin. The HF-pyridine method was used to identify all proteins located in the cell wall, including glycosylphosphatidylinositol (GPI)-anchored proteins, whereas trypsin shaving was used to identify proteins that contain surface-exposed, trypsin-accessible regions. A total of 148 proteins were identified, including 116 proteins by HF-pyridine extraction and 48 proteins by trypsin shaving, with 15 common in both data sets (Table 1; see also  Table S1 in the supplemental material). A total of 43 proteins contained a signal peptide for secretion (SignalP 4.1 Server [21]), 8 proteins contained predicted transmembrane helices (TMHMM Server v. 2.0 [22,23]), and 8 proteins contained a GPI-anchored attachment signal (according to the data in reference 24). As expected, the RodA surface hydrophobin was found to be the most abundant protein by both approaches. Along with known cell wall proteins, we also elucidated multiple proteins with unknown functions. The most abundant of these uncharacterized proteins was the conserved hypothetical protein Afu1g13670 denoted CcpA ( Table 1).
The levels of CcpA amino acid sequence coverage were 66.3% and 6.8% in the HF-pyridine and trypsin-shaving data sets, respectively (Fig. 1A). Only two peptides were identified by trypsin shaving, beginning at two consecutive lysine residues near the C-terminal end of the protein (KKASNPADSLGLGELTKVLGFR), suggesting that only a small portion of the protein is surface exposed. The ccpA gene is 757 bp in length, contains a single 48-bp intron at nucleotide 16, and encodes a predicted protein of 235 amino acids (deduced molecular mass, 25.7 kDa). CcpA contains a signal peptide for secretion and, interestingly, six conserved, invariant cysteine residues in the 15 organisms where it was readily identifiable (Fig. 1A; see also Fig. S1 in the supplemental material). The hydrophobicity plot of CcpA reveals an amphiphilic protein resembling RodA, with a hydrophobic N-terminal region and hydrophilic C terminus (Fig. 1B). CcpA is mainly found in Aspergillus species by BLAST analysis, and the closest homologues of A. fumigatus CcpA are in Neosartorya fischeri and A. udagawae (Fig. 1C). CcpA was deleted from wild-type A. fumigatus D141 (ΔccpA) and correspondingly complemented (ccpAc) (Fig. S2). Deletion of ccpA did not affect radial growth on solid media (Fig. S3). We also observed no significant differences in the germination rates of wild-type and knockout conidia produced on Aspergillus minimal medium (AMM) or malt agar plates (Fig. 1D). ΔccpA conidia showed no significant changes in comparison to wild-type conidia with respect to susceptibility to temperature stress (Fig. S3C), oxidative stress induced by H 2 O 2 (Fig. S3D), or cell wall/membrane stressors (Fig. S4A). Finally, scanning electron microscopy (SEM) of the surface of resting ΔccpA conidia revealed a rodlet layer identical to that of wild-type conidia, suggesting that deletion of ccpA does not significantly alter the surface characteristics of resting conidia. In CcpA is produced in phialides and is detectable in the cell wall of resting conidia. In order to confirm the localization of CcpA, a strain expressing CcpA fused to enhanced green fluorescent protein (eGFP) was generated ( Fig. S5A and B). When grown in AMM at 37°C, no fluorescence was detectable in hyphae (Fig. 2). However, during conidiogenesis, a strong fluorescence signal deriving from the conidiophore was readily visible. The phialides also emitted the eGFP signal, which strongly indicated expression of CcpA in these conidium-forming cells. High-resolution imaging confirmed FIG 2 CcpA is localized to the cell wall. CcpA_eGFP was cultivated in AMM at 37°C for analysis of resting conidia (0 h), swollen conidia (4 h), germinating conidia (6 h), hyphae (10 h), and conidiophore formation (24 h). Samples were analyzed by light and fluorescence microscopy. Scale bars ϭ 10 m.
Aspergillus fumigatus Immune Evasion ® that resting conidia displayed signal near the surface ( Fig. S5C and D). The eGFP fusion protein was also expressed in a ΔccpA knockout strain to ensure that the localization was comparable to the expression of the fusion in wild-type resting conidia ( Fig. S5E and F). During swelling of conidia, the eGFP signal appeared as scattered dots, which gradually accumulated in the vacuole, suggesting protein degradation (Fig. 2). In comparison, analysis of wild-type strain D141 as a control showed no fluorescent eGFP signal (Fig. S5G).
CcpA is necessary for a normal surface proteome. The surface localization and abundance of CcpA suggested a possible function in cell wall organization, and yet we observed no obvious cell surface differences in the knockout by SEM. Nor did we observe any alterations in the levels of mannose moieties of glycoproteins, chitin, or ␤-1,3-glucan in the cell wall of the knockout using concanavalin A-fluorescein isothiocyanate (concanavalin A-FITC), wheat germ agglutinin-FITC, or dectin-1 labeling, respectively (data not shown; the methods used were as described previously in reference 25). We also observed CcpA on the surface of both ΔrodA conidia lacking the rodlet layer and pksP mutant conidia lacking the melanin layer by trypsin shaving, suggesting that CcpA surface localization is likely not coupled to these conidial structures (M. G. Blango, T. Krüger, O. Kniemeyer, and A. A. Brakhage, unpublished data). To determine whether newly accessible surface proteins might be exposed on the ΔccpA conidial surface, we performed cell surface trypsin shaving of resting and swollen conidia (5 h). LC-MS/MS analysis of resting conidia from both malt and AMM agar plates revealed an increase in the number of identified proteins unique to the ΔccpA strain ( Fig. S6A and B). Ten proteins were found exclusively on the surface of resting ΔccpA conidia independently of the growth medium compared to the wild-type and complemented strains (Table S1). In swollen conidia, knockout spores grown on AMM agar displayed a similar phenotype, with a higher number of previously undetected proteins on the surface (Fig. S6C). Swollen conidia derived from malt agar plates were comparable to those derived from the wild-type and complemented strains, with only 17 proteins unique to the ΔccpA knockout surface (Fig. S6D). The identification of newly exposed surface proteins of the ΔccpA strain suggests that CcpA might limit the availability of A. fumigatus cell surface epitopes that would otherwise trigger inflammation upon each new encounter in healthy patients.
CcpA reduces recognition by the innate immune system. To test the hypothesis that CcpA contributes to innate immune evasion, we incubated primary polymorphonuclear leukocytes (PMNs), which mediate protection against invasive aspergillosis, with swollen conidia of the wild-type, ΔccpA, and ccpAc strains. All three strains induced an oxidative burst in neutrophils over time ( Fig. 3A and B). Formation of reactive oxygen intermediates (ROI) was strongly induced by ΔccpA conidia after 60 min of coincubation compared with mock-infected PMNs (Fig. 3A). At this time point, ROI generation in response to wild-type and ccpAc conidia was detectable at only low levels and remained significantly lower than that seen in response to ΔccpA conidia at 120 min (Fig. 3B). In line with this, PMNs confronted with ΔccpA conidia secreted significantly more of the neutrophil chemoattractant interleukin-8 (IL-8) (Fig. 3C) (26), whereas PMNs incubated with either the wild-type or ccpAc strain released only basal levels of IL-8. These data suggest that CcpA limits recognition by PMNs.
We next analyzed swollen conidia of the ΔccpA strain for the capacity to activate human monocyte-derived dendritic cells (moDCs). Formalin-fixed and swollen conidia (swollen for 3, 4, and 5 h) were coincubated with moDCs for 18 h. Interestingly, incubation of moDCs with ΔccpA swollen conidia resulted in enhanced release of the proinflammatory cytokines tumor necrosis factor alpha (TNF-␣), MIP-3a, IL-23, and IL-18 as well as the anti-inflammatory cytokines IL-10 and IL-4 ( Fig. 3D to I). Taken together, these data suggest that CcpA on the surface of conidia limits recognition by both PMNs and DCs.
CcpA contributes to both cell damage and virulence. The capacity of A. fumigatus to damage pulmonary epithelial cells in vitro correlates with virulence in mouse models of IA (27)(28)(29). Therefore, we investigated the effects of the ΔccpA deletion on the capacity of A. fumigatus to damage A549 pulmonary epithelial cells. Using a 51 Cr release assay for cytotoxicity, the ΔccpA strain was observed to cause significantly less damage to epithelial cells than the wild type after 16 and 24 h of incubation, suggesting decreased pathogenicity (Fig. 4A). This result was observed despite comparable levels of endocytosis for each strain (Fig. 4B). The modified surface proteome, the enhanced capacity of the ΔccpA knockout to stimulate PMNs and moDCs, and the reduced capacity to damage epithelial cells suggested that CcpA might be required for virulence. To test this assumption, we compared the virulence of the ΔccpA strain with that of the wild-type and ccpAc strains in a nonneutropenic mouse model of IA. In this model, mice are immunosuppressed with cortisone acetate, and yet recruitment of neutrophils to the site of infection still occurs (30).
The virulence of the ΔccpA strain was significantly attenuated (P Ͻ 0.002) in this mouse model (Fig. 5A). Over the course of infection, several mice infected with the ΔccpA strain developed characteristic disease symptoms (signs of dyspnea, ruffled fur), but the majority of these animals recovered, which led to a final survival rate of 80% after 14 days. In contrast, by 6 days postinfection with wild-type conidia, all animals had succumbed to the infection. Infection with the complemented ccpAc strain resulted in 90% killing of animals. Histopathology revealed that the wild-type strain induced strong inflammation and extensive pulmonary fungal invasion, with destruction of bronchi and alveoli. In contrast, the lungs of mice infected with ΔccpA conidia showed minimal evidence of inflammation and no fungi, suggesting early clearance of the conidia ( To demonstrate that the attenuated virulence of the ΔccpA strain was due to enhanced clearance by immune effector cells, we tested virulence in neutropenic mice generated by cyclophosphamide treatment. In addition to neutropenia, these mice also exhibit increased lung inflammatory features and decreased regulatory T cell levels (31,32). The ΔccpA strain was as virulent as the wild-type strain in this mouse model (Fig. 6A). Histopathology indicated clear signs of invasive growth of hyphae in tissue for all strains (Fig. 6B to E). These results demonstrate that the ΔccpA strain grows normally in the mouse lung and that the drastically attenuated virulence of this strain in nonneutropenic mice is likely due to interaction with the innate immune system.
The CcpA protein elicits a normal immunogenic T cell response. To test whether CcpA might play an active role in suppression of the adaptive immune response, we studied the reactivity of moDCs and recall T cell responses of healthy human donors against a recombinant CcpA protein purified as described in Materials and Methods (Fig. S8A). Following in vitro stimulation with recombinant CcpA, moDCs upregulated various costimulatory molecules (CD80, CD83, CD86, and CD40) and antigen-presenting molecules (HLA-ABC and HLD-DR) as well as proinflammatory cytokines compared to unstimulated controls ( Fig. S8B and C), indicating that recombinant CcpA protein has a boosting effect on the immune system via activated DCs.
To confirm this finding, we tested whether CcpA induces a specific immune response during natural exposure. Humans are continuously exposed to A. fumigatus; thus, in vivo primed T cells specific for immunogenic A. fumigatus proteins are readily detectable in the blood (33,34). Enrichment of antigen-reactive CD154 ϩ T cells (Tcons) following ex vivo CcpA stimulation of human blood cells allowed identification of in vivo primed CcpA-specific T cells according to cell surface phenotype and effector cytokine expression results. Specific T cells against CcpA could be clearly detected in healthy donors, albeit at a lower frequency than that seen following stimulation with wholeconidium extracts (Fig. 7A). However, compared to total A. fumigatus-specific T cells, the CcpA-specific T cells showed increased frequencies of memory cells (CD45RA Ϫ ) ( Fig. 7B and C) and inflammatory effector cytokine-expressing cells (Fig. 7D) as follows: for IFN-␥, a mean of 35.1% and a range of 7.9% to 78.9%; for IL-17A, a mean of 6.2% and a range of 0.53% to 20.4%). However, the level of expression of the anti-inflammatory cytokine IL-10 was reduced (mean, 2.68%; range, 0.55% to 6.59%) (Fig. 7D). These data identify an in vivo primed phenotype of CcpA-specific human T cells, suggesting that within the A. fumigatus proteome CcpA belongs to the group of immunogenic proteins (34). Overall, these data indicate that the CcpA protein has no protein-intrinsic immunosuppressive activity but rather seems to exert its function by limiting the recognition of resting conidia by altering the accessibility of the conidial surface, which, in the context of the immunosuppressed host, serves as a growth advantage for the fungus.

DISCUSSION
Proteins on the surface of conidia are likely to interact with the human immune system upon inhalation. In recent years, several studies have enlarged the list of A. fumigatus conidial surface proteins. Asif et al. treated intact resting A. fumigatus  (36). Here, we provide the most extensive overview to date of the conidial surface proteome, with 116 proteins identified by HF-pyridine extraction and 47 proteins identified by trypsin shaving, for a total of 148 unique cell wall proteins. We detected 22/25 of the proteins found by Asif et al. and 33/52 of the proteins found by Suh et al. in the cell wall of resting conidia. By comparing proteins identified by two complementary techniques, we were able to assess several aspects of the cell wall proteome. The HF-pyridine extraction technique releases cell wall and GPI-anchored proteins and uncovers the most abundant proteins in the cell wall, whereas the trypsin-shaving approach assesses the surface-exposed fraction of proteins that are trypsin accessible. Across all strains and conditions tested in the manuscript, including ΔccpA knockout conidia, we observed 701 unique proteins associated with the cell wall, including many known cell wall-associated proteins but also a number of unknown proteins (see Table S1 in the supplemental material) (35,36).
A large fraction of the identified proteins had no assigned biological function. The most abundant of these proteins was CcpA. Although this protein has a signal peptide, its amino acid sequence contains no other domain suggestive of cell wall localization. However, it was previously reported that CcpA is part of the conidial surface proteome (35). Localization studies performed with an eGFP-fusion protein revealed that CcpA was produced only in specific cell types, in the phialide layer formed on the premature conidiophore vesicle, as well as along the surface of conidiophores and resting conidia, consistent with the finding that CcpA is a resting-conidium-enriched protein (36). These findings are consistent with recent observations that ccpA appears to be regulated by the three transcription factors central to asexual development, BrlA, WetA, and AbaA (37), and that it contains a conserved, canonical AbaA binding site, with motif "CATTCC" (MEME analysis; reference 38 and data not shown). Due to the high abundance of CcpA in the cell wall and the cysteine-rich amino acid sequence, one may speculate that CcpA acts as a scaffold for the conidial hydrophobin layer. However, SEM pictures revealed that the ΔccpA knockout displays organized and compact rodlet structures over the entire resting conidial surface, similarly to wild-type conidia. Nevertheless, its amphiphilic character supports the idea of self-assembly through aggregation and a potential structural role in the organization of the conidial cell wall.
Intriguingly, proteomics analysis revealed new trypsin-accessible, surface-exposed proteins on ΔccpA resting conidia that could potentially contribute to its enhanced recognition and/or clearance. These phenotypes were also observed on the surface of swollen conidia from AMM agar plates, but not malt agar plates, perhaps suggesting a role for CcpA in modulating the surface environment under conditions of nutrient limitation or stress. The altered protein composition of the conidial surface offers compelling evidence for a role of CcpA in masking conidial surface antigens that could otherwise trigger an immune response in the host. It is not clear if these newly exposed proteins are actively hidden or if they just inherit the space made available by a missing CcpA protein. Although the presented evidence strongly suggests that CcpA limits immune recognition by PMNs and DCs, it is also formally possible that CcpA contributes indirectly to these fungal phenotypes.
The abundant surface localization and altered surface proteome of the CcpA knockout led us to test the contribution of CcpA to virulence. The ΔccpA knockout stimulated greater proinflammatory cytokine secretion and reactive oxygen species (ROS) production by PMNs and cell surface marker expression by moDCs. In nonneutropenic mice immunosuppressed with cortisone acetate, the ΔccpA strain showed significantly attenuated virulence. Interestingly, the knockout strain exhibited virulence comparable to that of the wild-type strain in cyclophosphamide-treated mice, with the mice succumbing to invasive fungal growth (30,39). In addition to neutropenia, these mice also exhibited other alterations to the normal immune response, including increased infiltration and abundance of lung eosinophils, decreased regulatory T cell levels, and loss of some subsets of fast-growing epithelial cells, among others (31,32,40). In addition to showing the importance of CcpA in promoting maximal virulence, this experiment also demonstrated that the ΔccpA deletion strain grows normally in an immunocompromised host, suggesting that reduced virulence in nonneutropenic mice was not solely due to a generalized growth defect.
In cell culture experiments, recombinant CcpA does not appear to be toxic to host cells (data not shown), consistent with our experiments in human T cells. In particular, our analysis of the in vivo immunogenicity of the purified CcpA protein suggests that the protein triggers a normal adaptive immune response upon natural exposure in healthy humans. We see no evidence for a protein-intrinsic immunosuppressive function or immunologic inertness. Thus, it seems that the presence of the protein on resting conidia is important for a normal surface structure. In the absence of CcpA, other proteins become more accessible on the conidial surface, leading to increased immune recognition and activation, in particular, by neutrophilic granulocytes. Although it requires further study, we hypothesize that CcpA serves as a major structural and stealth protein on the A. fumigatus conidial surface to prevent immune recognition by the innate immune system of the host.

MATERIALS AND METHODS
Strains and cultivation conditions. All strains used in this study are listed in Table S2 in the supplemental material. The clinical isolate D141 (41) served as the wild-type progenitor for all A. fumigatus strains constructed. Unless otherwise stated, A. fumigatus was cultivated at 37°C on Aspergillus minimal medium (AMM) agar plates or in AMM containing 50 mM glucose and 70 mM NaNO 3 as previously described (42,43). Media were supplemented with 0.1 g/ml pyrithiamine (Sigma-Aldrich) or 150 g/ml hygromycin (InvivoGen) when required. Conidia were harvested in sterile water. Escherichia coli strains were cultivated on LB agar plates or in LB supplemented with 50 g/ml kanamycin sulfate (AppliChem) at 37°C.
HF-pyridine extraction of conidial surface proteins. Proteins were extracted as previously described (44). A. fumigatus ATCC 46645 conidia grown for 7 days on malt agar were harvested, washed with sterile water, frozen in liquid nitrogen, and lyophilized. Lyophilized conidia (100 mg) were treated with 750 l hydrogen-fluoride (HF)-pyridine (Sigma-Aldrich) and neutralized with 1 M ice-cold Tris base, and proteins were precipitated with trichloroacetic acid (TCA)-acetone. After centrifugation for 20 min at 1,700 ϫ g at 4°C, the pellet was rinsed twice in ice-cold acetone-1 mg/ml dithiothreitol (DTT) before drying was performed for 15 min at room temperature. The extracted proteins were resuspended in 4 M urea-100 mM ammonium bicarbonate and sonicated for 20 min. After centrifugation at 20,800 ϫ g for 20 min at 16°C, protein concentration was determined using the Bio-Rad protein assay (Bio-Rad) (45).
Trypsin shaving of surface-exposed peptides. Seven-day-old resting and swollen conidia (5 h in RPMI 1640) from malt agar plates were washed twice with 25 mM ammonium bicarbonate and collected by centrifugation (1,800 ϫ g for 10 min). Samples were treated with 5 g trypsin (Serva) for 5 min at 37°C. Spores were separated from cleaved peptides using a 0.2 m-pore-size cellulose acetate filter (Sartorius) followed by inhibition of trypsin with formic acid (Sigma-Aldrich). Peptides were dried using a SpeedVac concentrator (Thermo-Fisher), resuspended in 25 l of 2% (vol/vol) acetonitrile (ACN)-0.05% (vol/vol) trifluoroacetic acid, and centrifuged for 15 min through a 10-kDa-cutoff microcentrifuge column (VWR). Due to continuous optimization of our sample preparation, the samples for swollen conidia only were instead passed through a 0.22-m-pore-size Spin-X cellulose acetate spin filter (Corning Costar), and the trypsin peptides were removed in silico.
Database search results were further processed with Scaffold version 3.0 (Proteome software) using the following parameters: minimum number of unique peptides ϭ 2, minimum peptide identification probability ϭ 95% (47). Normalized spectral counts were determined to assess relative protein abundance levels. Proteins were considered when at least two spectral counts were measured in two of three biological replicates.
LC-MS/MS analysis of trypsin-shaved surface peptides. Proteomics analysis was performed on an Ultimate 3000 RSLC nanoLC instrument coupled to a QExactive HF mass spectrometer (Thermo Fisher Scientific Company). Tryptic peptides were trapped for 4 min on an Acclaim PepMap 100 column (2 cm by 75 m, 3-m pore size) at a flow rate of 5 l/min. Subsequently, the peptides were separated on an Acclaim PepMap column ( , 100 ms). Fragment ions generated in the higher-energy collisional dissociation (HCD) cell at 30% normalized collision energy were scanned using N 2 (R, 15,000 FWHM; AGC target, 2e5; maximum IT, 100 ms) in a data-dependent manner (dynamic exclusion, 30 s).
Database search and data analysis of trypsin-shaved surface peptides. MS/MS data were searched against the A. fumigatus Af293 database of the Aspergillus Genome Database (AspGD) using Proteome discoverer 1.4 and the algorithms of Mascot 2.4.1, Sequest HT, and MS Amanda. 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. The dynamic modification was oxidation of Met. At least 2 peptides per protein and a strict target false-discovery rate (FDR) of Ͻ1% were required for positive protein hits.
Manipulation of DNA, Southern blotting. Manipulation of DNA was carried out according to standard procedures (48). Sequence information was obtained from the Aspergillus Genome Database (AspGD; http://www.aspergillusgenome.org) (49,50). Chromosomal DNA of A. fumigatus was isolated using a MasterPure Yeast DNA purification kit (Epicentre Biotechnologies). For Southern blot analysis, chromosomal DNA of A. fumigatus was digested with the indicated restriction enzymes (New England Biolabs). DNA fragments were separated in an agarose gel and blotted onto Hybond Nϩ nylon membranes (GE Healthcare Bio-Sciences). Labeling of DNA probes, hybridization, and detection of DNA-DNA hybrids were performed using digoxigenin (DIG) labeling mix, DIG Easy Hyb, and a CDP-Star ready-to-use kit (Roche Applied Science), respectively, according to the manufacturer's recommendations.
Genetic manipulation of A. fumigatus. Plasmids are listed in Table S2 and oligonucleotide sequences in Table S2. Plasmid pΔccpA was generated using plasmid pSK397 (51,52). pΔccpA contained the hygromycin B phosphotransferase gene (hph) under the control of the gpdA promoter and the trpC terminator of Aspergillus nidulans (see Fig. S1A in the supplemental material). The ccpA locus of A. fumigatus was targeted with a construct containing 1.2-kb flanking regions (5=-ccpA and 3=-ccpA). Oligonucleotides ccpA-1 to ccpA-8 encoded restriction sites at the 5= end for subsequent cloning steps and were used to amplify 5=-ccpA and 3=-ccpA by nested PCR. 5=-ccpA and 3=-ccpA were digested with NotI/XmaI and EcoRI/PacI, respectively, and were ligated end to end into pBluescript II S.K.(ϩ)_PacI (53). The hygromycin-resistance cassette was excised with SfiI from pSK397 and cloned into pBluescript to yield pΔccpA.
To complement strain ΔccpA, plasmid pccpAc was designed. ccpA and its 5= and 3= flanking regions were PCR amplified using primers v831/sv832_ptrA_re and sv833_ptrA_fw/sv834. The pyrithiamine resistance cassette was amplified from pSK275 with primers sv197/sv198. pccpAc was generated using a GeneArt seamless cloning and assembly kit (Invitrogen) according to the manufacturer's instructions (Fig. S1B).
For localization studies, plasmid pUC_GH_natpccpA_egfp was generated. Primers natpccpA_fw_Acc65I and ccpA_re_XmaI encoded additional 5= restriction sites and were used to amplify ccpA and an upstream 1,602-bp fragment comprising the native promoter. The amplified natpccpA DNA fragment was ligated in frame with egfp into pUC_GH to yield plasmid pUC_GH_natpccpA_egfp. Prior to transformation, the insertions of plasmids pΔccpA and pccpAc were excised by NotI/PacI and HpaI, respectively. Knockouts of A. fumigatus were generated by homologous recombination following the transformation of protoplasts (54). pUC_GH_natpccpA_egfp was incorporated into the genome of A. fumigatus D141 through ectopic integration. To study the functionality of the CcpA-eGFP fusion protein, the ΔccpA mutant was complemented with a ccpA-egfp fusion construct. For this purpose, plasmid pSK275_natpccpA_egfp was generated. The corresponding oligonucleotides nosT_rev_KpnI and ccpAp_for_KpnI were used with Flash-Phusion master mix (Thermo Scientific) to amplify the ccpAp-ccpA-egfp cassette from plasmid pUC_GH_natpccpA_egfp. The 3.4-kb PCR product was subcloned into pJet2.1 following the protocol of Aspergillus fumigatus Immune Evasion ® September/October 2018 Volume 9 Issue 5 e01557-18 mbio.asm.org 13 a CloneJET PCR cloning kit (Thermo Scientific). Then, the construct was inserted as a KpnI fragment into plasmid pSK275. Plasmid pSK275_natpccpA_egfp was used to transform protoplasts of the ΔccpA mutant. Pyrithiamine-resistant transformants were screened for fluorescence, and the complete ccpA-egfp sequence was verified by PCR with oligonucleotides nosT_rev and ccpAp_for. Production of recombinant CcpA protein in E. coli. A truncated version of CcpA lacking the 22-amino-acid secretion signal and 17 amino acids of the C terminus was produced. For this purpose, a synthetic gene with optimized codon usage for E. coli (Invitrogen, Life Technologies) was designed. By using oligonucleotides ccpA_23BamHIfw/ccpA_218HindIIIre, ccpA_23-218 was amplified from pMAT_ccpA_23-218. The ccpA_23-218 amplified PCR product was cloned into the BamHI/HindIII sites of plasmid pET28aH6TEV to give plasmid pET28aHTccpA_23-218, which was used to transform E. coli strain BL21(DE3). His 6 -ccpA-23-218 was produced by autoinduction in E. coli BL21(DE3) cells grown at 25°C in 1 liter of ZYP-5052 containing 0.5% (vol/vol) glycerol, 0.05% (wt/vol) glucose, 0.2% (wt/vol) ␣-lactose, and appropriate supplements (55). Cells (10 g by wet weight) were collected by centrifugation, resuspended in 100 ml lysis buffer, and homogenized using an Emulsiflex C5 high-pressure homogenizer (Avestin). Urea (8 M) was directly added to the lysate and dissolved. The fusion protein was isolated from clarified supernatant using a 5 ml HiTrap Talon crude column (GE Healthcare). Desalting and buffer exchange of His 6 -Afu1g13670_23-218 were performed using a 53-ml HiPrep 26/10 desalting column (GE Healthcare) by elution with 50 mM HEPES (pH 8.0)-300 mM NaCl.
Growth assays. Growth tests were performed on malt agar and on AMM agar plates containing either 50 mM glucose or 1% (wt/vol) peptone as the sole carbon source. Conidia (1 ϫ 10 3 in 5 l) were centrally spotted on agar plates in three technical replicates. The plates were incubated for 96 h, and radial growth was measured every 24 h.
The rate of germination was determined in RPMI media for spores collected from AMM or malt agar plates incubated at 37°C for 5 days. At 5, 6, 7, 8, and 9 h after the initiation of germination, an aliquot of spores was counted to determine the ratio of spores undergoing germination.
To analyze the susceptibility of conidia to oxidative stress, 1 ϫ 10 5 spores were treated with 0, 0.2, 0.4, or 0.6 M H 2 O 2 in a total volume of 1 ml for 30 min at room temperature. For testing the susceptibility of conidia at different temperatures, 1 ϫ 10 5 spores were incubated at Ϫ80°C, 22°C, 37°C, or 60°C for 1 h. After treatment, spore suspensions were diluted in water containing 0.001% (vol/vol) Tween 80 to a final concentration of 1 ϫ 10 3 /ml. A 100-l volume of each sample was plated on Sabouraud agar. After 24 h at 37°C, CFUs were counted and compared to an input that was plated from the initial dilution as described above.
Microscopy. For fluorescence and light microscopic analysis, 50 l AMM on glass coverslips in a wet chamber was inoculated with 2 ϫ 10 4 conidia. Samples were analyzed after 0, 4, 6, 10, and 24 h using a Zeiss Axio Imager.M2 (Zeiss). Images were taken with an AxioCam MRm and analyzed by the use of AxioVision SE64 Rel. 4.9.1 imaging software (Zeiss). For confocal scanning laser microscopy, a Zeiss LSM 780 instrument was used along with an Airyscan detector where noted. Images were processed using the ZEN software package from Zeiss.
For scanning electron microscopy (SEM) analysis, resting conidia from mycelia grown on AMM agar plates for 5 days were collected using an electrically conductive and adhesive tag (Leit-Tab; Plano GmbH). Samples for resting conidia were fixed for 24 h in a desiccator containing a solution of 25% (vol/vol) glutaraldehyde, whereas swollen conidia were fixed in 2.5% (vol/vol) glutaraldehyde. Further preparation and scanning electron microscopy were carried out as previously described (56).
For interaction studies, conidia were harvested after 5 days of growth on malt agar plates. Swelling was induced by incubating conidia in RPMI medium (Lonza) for 3, 4, or 5 h at 37°C. Fixation was achieved using 3% (vol/vol) formaldehyde for 1 h. Fixed conidia were subsequently pelleted by centrifugation. Pellets were washed with sterile water and RPMI medium containing 10% (vol/vol) heat-inactivated fetal calf serum (FCS) prior to stimulation.
Pulmonary epithelial cell damage assay. The A549 type II pneumocyte cell line was cultivated in F-12 K medium (American Type Culture Collection) containing 10% (vol/vol) fetal bovine serum (FBS) (Gemini Bio-Products), streptomycin, and penicillin (Irvine Scientific) in 5% (vol/vol) CO 2 at 37°C. Epithelial cell damage was measured using a standard 51 Cr release assay (29). Briefly, A549 epithelial cells were grown to 95% confluence in a 24-well tissue culture plate and loaded with 51 Cr (ICN Biomedicals). After removal of the unincorporated 51 Cr by rinsing, epithelial cells were infected with 5 ϫ 10 5 conidia in 1 ml of F-12 K medium per well. After incubation at 37°C in 5% (vol/vol) CO 2 for 16 and 24 h, the medium covering the cells was collected. The cells were lysed with 6 N NaOH, and the wells were rinsed with Radiac wash (Biodex Medical Systems, Inc.). The lysate and rinses were combined, and the amount of 51 Cr in the samples was determined by gamma counting. To measure the spontaneous release of 51 Cr, uninfected A549 cells exposed to medium alone were processed in parallel. After adjusting for well-to-well differences in the incorporation of 51 Cr, the percentage of specific release of 51 Cr was calculated using the following formula: (experimental release Ϫ spontaneous release)/(total incorporation Ϫ spontaneous release).
A549 endocytosis assay. The epithelial cell endocytosis of the wild-type, ΔccpA, and ccpAc strains was measured using a differential fluorescence assay as described previously (60). The three strains were transformed with a GFP expression plasmid containing the ble phleomycin resistance marker (GFP-Phleo) (61). Next, 10 7 conidia of each strain were added to 20 ml Sabouraud dextrose broth in a Petri dish and incubated at 37°C for 4.5 h to produce swollen conidia. A549 pulmonary epithelial cells grown on fibronectin-coated glass coverslips in a 24-well tissue culture plate were infected with 10 5 swollen conidia. After incubation for 3.5 h, the cells were washed with warm Hanks' balanced salt solution (HBSS) in a standardized manner to remove nonadherent fungi and then fixed with 4% (vol/vol) paraformaldehyde for 15 min. The noninternalized organisms were sequentially stained with a polyclonal rabbit anti-A. fumigatus serum (Meridian Life Science, Inc.) and anti-rabbit antibody (Ab) conjugated with Alexa Fluor 568 (Life Technologies). Coverslips were mounted inverted on a microscope slide and observed under conditions of epifluorescence. The number of organisms endocytosed by host cells was determined by subtracting the number of noninternalized organisms (with red fluorescence) from the total number of organisms (with green fluorescence). At least 100 organisms were counted on each coverslip.
Mouse infection models. Established murine models for invasive pulmonary aspergillosis were used for virulence studies (62,63). Female outbred CD-1 mice (Charles River) (18 to 20 g, 6 to 8 weeks old) were housed under standard conditions in individually ventilated cages and fed with normal mouse chow and water ad libitum.
Cortisone acetate (nonneutropenic) model. Mice were immunosuppressed with two single doses of 25 mg cortisone acetate (Sigma-Aldrich), which were injected intraperitoneally 3 days before and immediately prior to infection (day 0). Mice were anesthetized as described above and intranasally infected with 1 ϫ 10 6 conidia in 20 l PBS. Anesthesia was terminated by subcutaneous injection of flumazenil, naloxone, and atipamezol. Infected animals were monitored at least twice daily and humanely sacrificed if moribund (defined by severe lethargy, severe dyspnea, hypothermia, or substantial weight loss). Infections were performed with a group of 10 mice for each tested strain. A control group of 5 mice was mock infected (with PBS). For histopathological analyses, lungs from sacrificed animals were removed, fixed in formalin, and embedded in paraffin according to standard protocols. Sections (4 m) were treated with periodic acid-Schiff stain (PAS) using standard protocols. The sections were analyzed with a Zeiss Axio Imager M2 microscope (Zeiss). Images were taken with an AxioCam 105 color microscope camera and analyzed by the use of AxioVision SE64 rel. 4.9.1 imaging software (Zeiss).
Statistical analysis. Survival data were plotted as Kaplan-Meier curves and statistically analyzed by a log rank test using GraphPad Prism software 5.0 (GraphPad Software). The Student's t test was used for significance testing of two groups. Differences between the groups were considered significant at a P value of Յ0.05 or Յ0.01. Throughout the article, significance is denoted as follows: *, P ϭ Ͻ0.05; **, P ϭ Ͻ0.01, ***, P ϭ Ͻ0.001; ns, nonsignificant.
Ethics statement. For PMN experiments, human peripheral blood was collected from healthy volunteers after written informed consent was provided. The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the University Hospital Jena (permit number 273-12/09). For DC experiments, ethical approval was obtained by the Ethical Committee of the University Hospital, Würzburg, for the use of blood of healthy donors (approval 34/15). Written informed consent was provided by all study participants. For T cell experiments, all donors gave consent and all protocols were approved by the Ethics Committee Charité (EA1/149/12). Mice were cared for in accordance with the principles outlined by the European Convention for the Protection of Vertebrate Animals Used for Experimental and Other Scientific Purposes (European Treaty Series, number 123; http://conventions.coe.int/Treaty/en/Treaties/Html/123.htm). All animal experiments were in compliance with the German animal protection law and were approved by the responsible Federal State authority "Thüringer Landesamt für Verbraucherschutz" and ethics committee "Beratende Komission nach §15 Abs. 1 Tierschutzgesetz" with permit number 03-001/12.

ACKNOWLEDGMENTS
We gratefully acknowledge Silke Steinbach, Sylke Fricke, Carmen Schult, and Maria Pötsch for their excellent technical assistance and Jenny Kirsch and Toralf Kaiser from the DRFZ Flow cytometry Core Facility. We also acknowledge Marie Röcker for her work in optimizing the trypsin-shaving proteomics approach, Kai Naumann for his help with analyzing the LC-MS data, and Hella Schmidt for her help with microscopy. This work was supported by the Deutsche Forschungsgemeinschaft (DFG; http:// www.dfg.de/en/)-funded Excellence Graduate School Jena School for Microbial Communication (JSMC; http://www.jsmc.uni-jena.de/) and the DFG-funded Collaborative Research Center/Transregio 124 "Human-pathogenic fungi and their human host-Networks of interaction-FungiNet" (http://www.funginet.de) (projects A1, A2, C3, and Z2). M.G.B. was funded by EXASENS project 13N13861, and M.S. was funded by InfectControl 2020 project 03ZZ0803A by the Federal Ministry of Education and Research (BMBF; https://www.bmbf.de/), Germany. A.S. and P.B. were supported by BMBF grants for InfectControl 2020 projects ART4Fun Fkz 03ZZ0813A and DIAT Fkz 03ZZ0827A. F.S. was supported by the Leibniz Science Campus InfectoOptics. S.G.F. was supported by grants R01AI073829 and R56AI111836 from the National Institutes of Health, USA (NIH; https://www.nih.gov/). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.