Immunoproteome of Aspergillus fumigatus Using Sera of Patients with Invasive Aspergillosis

Invasive aspergillosis is a life-threatening lung or systemic infection caused by the opportunistic mold Aspergillus fumigatus. The disease affects mainly immunocompromised hosts, and patients with hematological malignances or who have been submitted to stem cell transplantation are at high risk. Despite the current use of Platelia™ Aspergillus as a diagnostic test, the early diagnosis of invasive aspergillosis remains a major challenge in improving the prognosis of the disease. In this study, we used an immunoproteomic approach to identify proteins that could be putative candidates for the early diagnosis of invasive aspergillosis. Antigenic proteins expressed in the first steps of A. fumigatus germination occurring in a human host were revealed using 2-D Western immunoblots with the serum of patients who had previously been classified as probable and proven for invasive aspergillosis. Forty antigenic proteins were identified using mass spectrometry (MS/MS). A BLAST analysis revealed that two of these proteins showed low homology with proteins of either the human host or etiological agents of other invasive fungal infections. To our knowledge, this is the first report describing specific antigenic proteins of A. fumigatus germlings that are recognized by sera of patients with confirmed invasive aspergillosis who were from two separate hospital units.


Introduction
Invasive aspergillosis is a life-threatening lung or systemic infection that primarily affects hematological patients under chemotherapy and hematopoietic stem cell transplant (HSCT) patients [1]. The infection is fatal in 30%-90% of the patients, including those given treatment [2]. The main etiological agent of invasive aspergillosis is the opportunistic mold Aspergillus fumigatus, which is responsible for 90% of aspergillus infections [3].
A confirmed diagnosis of invasive aspergillosis remains challenging and is frequently not achieved until necropsy. The isolation of aspergilli from cultures lacks sensitivity and, therefore, is ineffective for the diagnosis of invasive aspergillosis; blood cultures are rarely positive even in patients with confirmed invasive aspergillosis [4,5]. Moreover, the isolation of aspergilli in blood cultures or in sputum samples does not necessarily indicate the presence of the invasive disease. Positive results usually represent only colonization due the high colonization rate in immunocompromised patients; thus, false-positive results due environmental contamination are frequent [5,6].
The "gold standard" for the diagnosis of invasive aspergillosis remains histopathological examination or biopsy; however, this often requires invasive procedures to obtain tissue for the examination. In most cases, the aggressiveness of the underlying disease, as well as the toxic effects of the hematological therapies, make this type of examination impossible in critically ill patients [3,7,8].
Currently, the routine techniques used for the diagnosis of invasive aspergillosis are computational tomography (CT) and the ELISA test for galactomannan (GM) (Platelia™ Aspergillus-BioRad, Hercules, CA, USA); these are considered along with microbiological findings and the clinical signs and symptoms of the patient [9,10]. The GM molecule is an immunodominant cell wall polysaccharide of Aspergillus and Penicillium species that is released during fungus growth [11,12]. Although it provides a fast serological result, the efficacy of the GM test remains controversial and varies depending on the clinic or health center, as previously reviewed by Xavier et al. [13]. False-positives have also been reported, for example, following treatment with a beta-lactam antibiotic; however, recent reports suggest that the new preparations of piperacillin-tazobactam do not test positive with galactomannan. Cross-reactions with fungi, such as Fusarium spp., Penicillium, Cladosporium and Histoplasma have also been reported [14][15][16][17]. The mean specificity of the test is 85% and the sensitivity varies from 29% to 100% [9,13].
The difficulties in reaching an early and precise diagnosis are also true for other invasive fungal infections. To define and classify the main invasive fungal infections in immunocompromised patients, the European Organization for Research and Treatment of Cancer/Invasive Fungal Infections Cooperative Group and the National Institute of Allergy and Infectious Diseases Mycoses Study Group (EORTC/MSG) Consensus Group created and revised the definitions for clinical and epidemiological research. According to the definitions, invasive fungal infections are classified as "proven", "probable", or "possible" [5,18]. Thus, there remains an urgent need to develop new diagnostic tools to prevent the onset of the disease.
The sequencing of the A. fumigatus genome and the advances in the proteomic field have made it feasible to study and identify putative candidates for the immunodiagnosis of invasive aspergillosis. Few antigens specific for allergic bronchopulmonary aspergillosis (ABPA), aspergilloma, and invasive aspergillosis are known and/or being evaluated for diagnosis [19]. Furthermore, some studies have already shown the potential of some proteins as biomarkers for the immunodiagnosis of invasive aspergillosis; however, none of these came to a clinical trial [20][21][22][23][24].
In this context, the aim of this study was to investigate the antigenic proteins revealed by patients' sera using cell wall extracts of A. fumigatus germlings in an attempt to find putative candidates for the diagnosis of invasive aspergillosis.

Western Immunoblots and Antigenic Proteins Identified
In recent decades, invasive fungal infections (IFI) have been considered the most important cause of morbidity and mortality in severely immunosuppressed patients. Although candidiasis remains the most frequent IFI in critically ill patients, aspergillosis and mucormycosis have also emerged as significant causes of morbidity and mortality. HSCT recipients and patients with prolonged neutropenia represent the main risk group for invasive aspergillosis [25]. In these patients, A. fumigatus is by far the most important etiological agent of invasive aspergillosis, especially in HSCT patients with acute leukemia (5% to 25%) and in some solid organ transplantation patients [3,7,26].
As mentioned previously, the actual diagnostic methods lack specificity and sensitivity for the early diagnosis of invasive aspergillosis. In this context, many efforts have been undertaken to identify new molecular tools that could reduce this difficulty. Immunoproteomic-based antigen identification is a convenient tool that is widely used to indicate putative candidates for the molecular diagnosis of fungal infections, including invasive aspergillosis [23,24,27,28]. Germlings are cells in an early stage of growth, and surface proteins in this morphotype may play an essential role in the fungal-host interaction [29,30]. In addition, the cell surface location of these proteins makes germlings more easily recognized by the host immune system [12]. Thus, proteins present in the A. fumigatus germling cell wall can represent important putative antigenic markers for the early diagnostic of invasive aspergillosis.
In this study, the antigenic profile of cell surface proteins of A. fumigatus germlings (GT 6 h ) were identified through an immunoproteomic approach. The 2-DE profile of the GT 6 h extract, obtained as previously described [31], is shown in Figure 1. All antigenic proteins identified in this study, as well as their molecular mass, isoelectric points and functions, are listed in Table 1. The Western immunoblot analysis using the distinct pools of human sera, which were typed following    A total of fourteen antigenic proteins were exclusively revealed by sera of patients with proven aspergillosis, as shown in Table 2 (grey lines). Among these, four proteins were also recognized by pool of sera classified as probable by the EORTC/MSG criteria. Five out of fourteen proteins were positively recognized by the pool of patients with proven aspergillosis from both Hospital 1 and 2. Some of these identified antigens had also been described in other reports based on assays with the sera of immunized rabbits, mice and of patients with the clinical suspicion of allergic bronchopulmonary aspergillosis [20,21,23,28,32]. To our knowledge, this work is the first to describe four antigens: eEF-3, eIF4A, cytochrome P450 and Ade1, which are putative candidates for diagnostic utility.

BLAST Analysis
The fourteen antigens revealed from the immunoproteome of the sera from patients with proven invasive aspergillosis (n = 12) were selected as putative candidates for the diagnosis of invasive aspergillosis. Their protein sequences were compared with human proteins via BLAST analyses to ensure their potential specificity for A. fumigatus and cross-reactivity with human proteins. Our results showed that only two antigenic proteins, cytochrome P450 and eEF-3, had no homology with human proteins.
As mentioned previously, the diagnosis of invasive aspergillosis can be confused with a range of other invasive fungal infections [33][34][35][36][37]. In this context, we also compared (via BLAST analysis) the sequences of the two above-described proteins with proteins of Rizophus spp. and other fungi from the Mucorales order; Penicillium spp., Paracoccidioides brasiliensis, Fusarium spp., and Paecilomyces spp., as described in the methodology section. The results shown in Table 3 indicate that both cytochrome P450 and eEF-3 can be putative markers for the selective diagnosis of A. fumigatus infections.  The cytochrome P450 superfamily is made up of monooxygenases that play key roles in a range of biochemical processes from catalysis to xenobiotic detox and degradation; cytochrome P450 is found in every living form [38]. In general, cytochrome P450 isoforms have being described as essential for the membrane ergosterol biosynthesis, and some isoforms are involved in the production of aflatoxin in A. parasiticus [39][40][41]. In A. fumigatus, triazole resistance is often related to mutations in a gene that encodes a cytochrome P450 isoform, the cyp51 gene [42][43][44]. Although the secondary structures of the proteins of the cytochrome P450 superfamily are well conserved, there is a low homology among the primary amino acid sequences of different species [45][46][47][48]. These data are consistent with the result of our BLAST analysis that shows the low homology of the identified A. fumigatus cytochrome P450 found in this study with proteins of other fungi ( Table 3). The cytochrome P450 identified in this study is predicted in the A. fumigatus genome but has no characterized function. To our knowledge, this is the first report showing the antigenic diagnostic potential of an A. fumigatus cytochrome P450.
The most promising antigen was the translation elongation factor eEF-3. This protein showed the lowest sequence homology in the BLAST analysis ( Table 3). The translation process functions in a series highly regulated steps that are catalyzed by the eukaryotic initiation factors [49]. In general, the process is highly conserved from bacteria to mammals: the eEF-1 is incumbent on delivering the aminoacyl-tRNA to the ribosomal A-site [50], and the eEF-2 has a translocase activity [51]. However, another factor is required in fungi (an ATPase factor, namely eEF3). This requirement is unique in fungi ribosomes. This fungal-specific protein is absent in mammalian cells and has already being described by our group as a putative drug target in A. fumigatus [27]. The eEF-3 is an ATPase of the ATP binding cassette (ABC) family member [52]. The majority of this superfamily's members are integral membrane transporters that are involved in the import or export of diverse substrates across lipid bilayers [53]. However, eEF-3 lacks the transmembrane domain because it is a soluble factor with two ABC domains arranged in tandem. One of these domains carries a unique chromodomain-like insertion that is hypothesized to play a significant role in its binding to the ribosome [54]. A recent study showed that mutations in the chromodomain-like insertion of eEF-3 resulted in reduced growth rate and slower translation elongation. These mutations also compromised the ribosome-stimulated ATPase activity of eEF3, strongly suggesting that it exerts an allosteric effect on the hydrolytic activity of eEF3 [55]. These features contributed to the overexpression of eEF-3 in the first steps of A. fumigatus filamentation (germlings), strengthening the hypothesis that this protein may be a good drug target [31].
Our previous studies showed that this protein was found to be overexpressed up to eight-fold on the surface of the germlings compared with mature A. fumigatus hyphae [31]. In this study, the eEF-3 factor was identified as an antigenic protein of A. fumigatus recognized by the sera of patients with proven invasive aspergillosis. Taken together, these observations strongly suggest that in addition to being a putative drug target, the identified A. fumigatus eEF-3 factor can also be a promising candidate for the diagnosis of invasive aspergillosis.

Fungal Strain and Culture Conditions
The A. fumigatus strain used in this study was AF293, which was originally isolated at autopsy from a patient with IPA and kindly provided by Dr. Scott Filler of Harbor-UCLA Medical Center, University of California, CA, USA.
A. fumigatus was first grown in Sabouraud Agar (Difco, Detroit, MI, USA) roux flask for 7 days at 37 °C. The conidia were than harvested using a cell scraper in the presence of PBS-Tween 20 (0.01%). This suspension was vacuum-filtered using a Büchner filler with a nylon membrane (Sefar Nitex 03-28/17, 7, Sefar Inc., Heiden, Switzerland) to remove hyphae fragments. A ratio of 10 7 conidia/mL was then incubated in Sabouraud Broth (Difco, Maryland, MD, USA) in a 500-mL flask on a shaker at 37 °C and 150 rpm for 6 h to obtain the conidia germlings.

Preparation of Germiling Conidia Protein Extract (GT 6 h )
Conidia germling cells were submitted to chemical extraction [56] using protein extraction buffer containing Tris-HCl 25 mM, DTT 2 mM, PMSF 1 mM and EDTA 5 mM, pH 8.5. The conidia germling cells were incubated with the protein extraction buffer in a ratio of 0.7 g of cells (wet weight) per 5 mL of buffer for 2 h at 4 °C under gentle agitation. The proteins extracted using this process were separated via centrifugation. The extract was precipitated with trichloroacetic acid/acetone [57] and re-suspended in rehydration buffer containing 7 M urea, 2 M thiourea and CHAPS 4%. The protein concentration was determined using the Bradford method (Bio-Rad, Hercules, CA, USA) according to the manufacturer's recommendations. The absence of membrane leakage and consequently intracellular proteins or material derived from dead cells, in this type of extraction have been previously described [27].

Patients and Control Subjects
All of the serum samples of patients were obtained with informed patient consent and the permission of the local human ethics committee. All serum samples were classified according to the EORTC/MSG criteria [18]. Three serum samples of patients clinically diagnosed as proven and thirteen serum samples of patients clinically diagnosed as probable were obtained from the Bone Marrow Transplant Center of the National Institute of Cancer (INCA-Brazil), henceforth referred to as Hospital 1. More information about the characteristics of the patients from Hospital 1 is shown in Table 4. Nine serum samples of patients classified as "proven" for invasive aspergillosis were obtained from the Hospital das Clínicas of the Faculty of Medicine from the University of São Paulo (USP-Brazil), henceforth referred to as Hospital 2. Serum samples from patients with other fungal infections viz. histoplasmosis (n = 1), fusariosis (n = 3), cryptococcosis (n = 1) and paracoccidioidomycosis (n = 1) were also provided by Hospital 2. These patients had also underlying diseases similar to those found in the aspergillosis cases. As a negative control, sera from six patients with underlying diseases similar to the aspergillosis cases, such as acute myeloid leukemia (n = 2), non-Hodgkin lymphoma (n = 2), multiple myeloma (n = 1) and myelodysplastic syndrome (n = 1), were also provided by Hospital 2. These patients did not receive antifungal treatment, presented no colonization by any fungal species and survived for at least 30 days. More information about the characteristics of the patients from Hospital 2 is shown in Table 5. The serum samples were pooled for the immunoproteome assays as follows: proven/hospital 1, proven/hospital 2, probable or other-mycosis.

Western Immunoblot
For the immunoblottings, the resolved proteins were transferred to nitrocellulose membranes using a Trans-Blot Cell system (Bio-Rad). The transblotted proteins on the membrane were checked with Ponceau, and each membrane was blocked with 5% skim milk solution in 50 mM Tris and 150 mM NaCl containing 0.1% of Tween-20 (TBS-T). Then, the membranes were washed with 1% skim milk solution in TBS-T and incubated separately with each primary antibody (pools of sera: proven/hospital 1, proven/hospital 2, probable, other-mycosis, control) diluted in TBS-T at a 1:500 ratio for two hours at 4 °C under gentle agitation. The membranes were washed with 1% fat free milk solution in TBS-T (as above) and incubated with the secondary antibody (anti-human IgG peroxidase conjugated) (Sigma Co., St Louis, MO, USA) diluted in TBS-T at a 1:1000 ratio for two hours at 4 °C under gentle agitation. After washing with TBS, the membranes were incubated with the ECL Prime Western Blotting Detection Reagent (GE Healthcare, Menlo Park, CA, USA) according to the manufacturer's recommendations, and the antigenic spots were visualized using a Molecular Imaging ChemiDoc XRS system (Bio-Rad, Hercules, CA, USA).

Protein Identification
Spots of interest were manually excised from the preparative 2-DE gels. These spots were destained, shrunk, vacuum-dried, as described elsewhere [27] and then, were incubated with 12.5 ng/μL sequencing grade trypsin (Promega, Madison, WI, USA) overnight at 37 °C. After digestion, the supernatants were separated and the peptides were extracted twice into 0.5% trifluoroacetic acid/50% acetonitrile and once into 100% acetonitrile. These extracts were pooled, and their volumes were vacuum-dried. The derived concentrated peptide suspension for each spot of interest was spotted on a MALDI target plate, mixed with a saturated solution of matrix α-cyano-4-hydroxytrans-cinnamic acid (Sigma Co., St Louis, MO, USA) and allowed to air-dry at room temperature. The samples were analyzed with a 5800 AB-SCIEX MALDI-TOF/TOF mass spectrometer (Applied Biosystems, Foster City, CA, USA) in automated mode. A MALDI MS spectrum was acquired from each spot (800 shots/spectrum), and 10 precursor peaks with a signal-to-noise ratio greater than 40 in at least two consecutive fractions were automatically selected for MS/MS analysis (4000 shots/spectrum). A collision energy of 1 keV was used with air as the collision gas. All mass spectra were externally calibrated using the mass standards kit for the 4700 proteomics analyzer (Applied Biosystems, Foster City, CA, USA). The spectra were searched against an in-house database constructed using "A. fumigatus" as the selection criteria in Protein Pilot software using the Paragon algorithm (Applied Biosystems, Foster City, CA, USA). The name of the ORF (open reading frame) from A. fumigatus was found in the UniProt (Universal Protein Resource) server using the UniProt Knowledge/Swiss-Prot database.

Homology Analysis
The sequences of the antigenic proteins were aligned and compared using the protein BLAST tool of the NCBI database (http://blast.ncbi.nlm.nih.gov). The sequences of the identified A. fumigatus proteins were compared with sequences of human proteins and with proteins from other microorganisms. The selected microorganisms for comparison in the BLAST analyses are the etiological agents of mycosis that can be confused (diagnostically) with invasive aspergillosis (Rizophus spp. and other fungi of the Mucorales order, Penicillium spp., Paracoccidioides brasilienisis, Fusarium spp. and Paecilomyces spp.). The proteins with identity values lower than 40% and E-values higher than 1 × 10 −50 were identified to have no homology.

Conclusions
Two antigenic proteins of A. fumigatus are described in this work as putative candidates for the immunodiagnostic of invasive aspergillosis: cytochrome P450 and eEF-3. These proteins presented no homology with human proteins and low homology with etiological agents of other IFIs. Among these, the elongation factor eEF-3 identified in A. fumigatus germlings is the most promising candidate once it shows the lowest homology with proteins of other fungal species that cause infections, which could be misdiagnosed with invasive aspergillosis.