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 (). All animal experiments were in compliance with the German animal protection law and were approved by the responsible Federal State authority “Thüringer Landesamt für Lebensmittelsicherheit und Verbraucherschutz” and ethics committee “Beratende Kommission nach §15 Abs. 1 Tierschutzgesetz” with the permit Reg.-Nr. 03-001/12.

All A. fumigatus strains used in this study are listed in Table S1. A. fumigatus strain CEA17ΔakuB (da Silva Ferreira et al., 2006) was used to generate mutant strains ΔfhpA, ΔfhpB, ΔgnoA, ΔgnoA/ΔfhpA, ΔgnoA/ΔfhpB, and ΔfhpA/ΔfhpB. The strains were cultivated in Aspergillus minimal medium (AMM; Brakhage and Van den Brulle, 1995) containing 50 mM glucose and 70 mM NaNO₃ as sole carbon and nitrogen source, respectively, if not otherwise stated. Alternative nitrogen sources were ammonium tartrate (20 mM) or glutamine (20 mM).

Standard techniques for manipulation of DNA, including isolation of genomic A. fumigatus DNA and Southern blot analyses, were carried out as described earlier (Wartenberg et al., 2011). For RNA isolation and Northern blot analyses the protocols previously described (Bergmann et al., 2010) were followed. RNA samples were incubated with Turbo DNA-free^TM Kit (Ambion, USA) and transcribed in cDNA with the RevertAid^TM Premium First Strand cDNA Synthesis Kit according to the manufacturer’s protocol (Fermentas, Germany).

Transformation of A. fumigatus protoplasts was performed as described earlier (Weidner et al., 1998). For deletion of gnoA (AFUA_2G01040) the flanking regions were amplified from genomic A. fumigatus DNA with the primer pairs gnoA_5for/gnoA_ptrA_5rev and gnoA_3rev /gnoA_ptrA_3for. The pyrithiamine resistance cassette ptrA was amplified with primers ptrA_forII/ptrA_revII from plasmid pSK275 (Szewczyk and Krappmann, 2010). For assembly of the gnoA deletion construct, a 3-fragment PCR according to the procedure described previously (Scharf et al., 2011) was performed with primers gnoA_5for and gnoA_3rev. For generating the fhpB (AFUA_8G06080) deletion mutant the flanking regions were amplified with the primers fhpB_5for/fhpB_hph_5rev and fhpB_3rev/fhpB_hph_3for. Here, hygromycin resistance was used as selection marker and the hygromycin resistance cassette was amplified from plasmid pUChph (Liebmann et al., 2004) with the primer pair hph_for/hph_rev. Assembly of the fhpB deletion construct was achieved by 3-fragment PCR with primers fhpB_5for/fhpB_3rev. The deletion construct to create the ΔfhpA-mutant was generated by amplifying the flanking regions of fhpA using the primer pairs fhpA_1(XmaI)/fhpA_2(SfiI) and fhpA_3(SfiI)/fhpA_4(XhoI). The resulting DNA fragments were separately cloned into pCR2.1TOPO (Invitrogen, Germany) and subsequently isolated by restriction with XmaI/SfiI and SfiI/XhoI. The pyrithiamine resistance cassette was obtained by SfiI-restriction of plasmid pSK275. The fhpA DNA flanking fragments and the SfiI-digested ptrA gene were cloned into plasmid p123 (Spellig et al., 1996) restricted with XmaI/XhoI. The resulting plasmid was digested with XmaI and XhoI and the deletion cassette was used to transform A. fumigatus CEA17ΔakuB or ΔfhpB to create the ΔfhpA mutant and the ΔfhpA/ΔfhpB double mutant, respectively. For creating the double mutants ΔfhpA/ΔgnoA and ΔfhpB/ΔgnoA the respective gnoA deletion cassettes conferring pyrithiamine or hygromycin resistance were employed to transform the mutants ΔfhpA and ΔfhpB, respectively. The gnoA deletion cassette containing the hygromycin resistance cassette was generated by 3-fragment PCR as described above employing primers gnoA_5for, gnoA_hph_5rev, gnoA_3rev, and gnoA_hph_3for. For complementation of the ΔgnoA mutant and construction of a GnoA-eGFP strain, the gnoA gene including 1 kb promoter sequence, was amplified from genomic A. fumigatus DNA using primers gnoA_Acc65I_for and gnoA_XmaI_rev. The DNA fragment was inserted in vector pJet1.2/blunt (Fermentas, Germany). The gnoA gene was inserted as Acc65I and XmaI fragment in plasmid pUCGH resulting in pUCGH_gnoA-egfp that was used to transform A. fumigatus ΔgnoA. Complementation of the ΔfhpA mutant by construction of an FhpA-eGFP strain is described in Kroll et al. (2014). For generation of strain FhpB-eGFP containing the fhpB-egfp gene fusion under control of the constitutive otef promoter, the fhpB gene was amplified by PCR using primers FhpB_XmaI_rev and FhpB_XmaI_for_otef. The DNA fragment was ligated via XmaI into plasmid pTH1 (a derivative of pUCGH carrying the pyrithiamine resistance cassette instead of the hygromycin resistance cassette) resulting in plasmid pTH-fhpB-egfp that was used for transformation of A. fumigatus CEA10. All PCR reactions were performed using Phusion high-fidelity DNA polymerase (Finnzymes, Finland). All oligonucleotides used in this study are listed in Table S2.

Petri dishes with 10 ml bottom-agar were covered with a layer of 10 ml top-agar containing 1 × 10⁸ conidia of the respective strain. In the center, 100 μl of 45 mM DETA-NO (Sigma, Germany) were filled in a hole of 10 mm diameter.

1 × 10⁴ conidia of A. fumigatus were cultivated in 50 μl AMM on cover slips and incubated at 37^∘C in a wet chamber. The ratio of germinated to non-germinated conidia was determined by light microscopy by counting 100 conidia/germlings per time point and strain. The assay was repeated three times.

Fifty microliter AMM were inoculated with 1 × 10³ conidia of strain FhpA-eGFP or GnoA-eGFP on glass cover slips and incubated at 37^∘C. To analyze dormant conidia, conidia were inoculated in 50 μL H₂O. For staining mitochondria, conidia of strain FhpB-eGFP were cultivated in 48 well plates with cover slips for 16 h in AMM. Fresh AMM with 500 nM Mitotracker Deep Red (Invitrogen) was added and staining was done for 30 min at 37^∘C. Before fluorescence microscopy, the samples were washed three times with PBS. Samples were analyzed using a Leica DMI 4000B and a DM 4500B microscope (Leica Microsystems, Germany). Images were taken with a Leica DFC480 or a DFC 340FX camera and analyzed by Leica LAS V.3.7 software.

Alveolar mouse macrophages (MH-S, ATCC CRL-2019) were grown in 12-well plates at a concentration of 7.5 × 10⁵ cells per well over night at 37^∘C and 5% (v/v) CO₂ in RPMI 1640 supplemented with 10% (v/v) FCS and 50 μg ml^-1 gentamycin (PAA Laboratories, Germany). After washing with PBS the cells were co-incubated with 1.5 × 10⁶ A. fumigatus conidia for 20 min at 4^∘C and then incubated for 6 h at 37^∘C at 5% CO₂. The samples were centrifuged with 5,000 g for 5 min and resuspended in 1 ml ice-cold water and incubated for 10 min on ice. The samples were mixed and diluted 1:1000 with PBS/0.01% (v/v) Tween 20. 100 μl of the sample were plated on Sabouraud agar plates and colonies were counted after incubation for 24 h at 37^∘C. For determination of CFU, five technical replicates were analyzed.

Virulence of A. fumigatus mutant strains was tested in an established murine model for invasive pulmonary aspergillosis (Kupfahl et al., 2006; Schobel et al., 2010). Female CD-1 mice (Charles River, Germany) were immunosuppressed with cortisone acetate (25 mg per mouse intraperitoneally; Sigma-Aldrich, Germany) 3 days before and on the day of infection. Mice were anesthetized and intranasally infected with 20 μl of a fresh suspension containing 2 × 10⁵ conidia. A control group was mock-infected with PBS. To induce neutropenia in mice, cyclophosphamide (140 mg kg^-1; Sigma–Aldrich, Germany) was injected intraperitoneally on days -4, -1, 2, 5, 8, and 11, with an additional subcutaneous dose of cortisone acetate (200 mg kg^-1) on day -1. Mice were anesthetized and 3 × 10⁴ conidia in 20 μl PBS were applied intranasally.

The health status was monitored at least twice a day for 14 days and moribund animals, defined by severe dyspnoea, severe lethargy, or weight loss >20%, were sacrificed. Infections were performed with a group of 10 mice for each tested strain. Lungs from sacrificed animals were removed, and either stored in RNAlater (Qiagen, Germany) for RNA extraction or fixed in formalin and paraffin-embedded for histopathological analyses according to standard protocols. RNA isolation and first-strand cDNA synthesis from infected lungs was performed as described previously (Schmaler-Ripcke et al., 2009).

Virulence of the A. fumigatus ΔfhpA/ΔfhpB mutant was tested in a non vertebrate infection model using fresh larvae of the greater wax mothGalleria mellonella (Kavanagh and Reeves, 2004). For each strain 30 larvae (Quality Bugs, Linnich, Germany) were infected with 5 μl of an A. fumigatus spore suspension (5 × 10⁷ spores/ml in DPBS + 0.01% (v/v) Tween 20). Injection site was the lower right pseudopod. Larvae were checked for survival twice a day and accounted dead if nudging with forceps did not result in movement.

The Student’s t-test was used for significance testing of two groups. All significant differences are labeled with an asterisk (^∗p ≤ 0.05; ^∗∗p ≤ 0.01).

To elucidate the role of RNI in the pathogenicity of A. fumigatus, genes encoding putative RNI-detoxifying enzymes were identified. BLAST analyses revealed that the A. fumigatus genome contains two putative orthologs to the yeast flavohemoglobin YHB1, named FhpA (AFUA_4G03410) and FhpB (AFUA_8G06080). The FhpA and FhpB amino acid sequence revealed 40 and 36% identity, respectively, to YHB1. FhpA and FhpB share 48% identical amino acids. As a third protein potentially involved in NO detoxification, GnoA (AFUA_2G01040) was identified as ortholog (73% identity) of the nitrosoglutathione reductase Gno1 previously characterized in C. neoformans (Zhou et al., 2009).

Single and double deletion mutants of FhpA, FhpB, and GnoA encoding genes were generated (supplemental figures S1–S7). To test the sensitivity of A. fumigatus against nitrosative stress, an inhibition zone assay using DETA-NO as NO donor was performed (Figure 2A). Strains ΔfhpA and ΔgnoA showed increased sensitivity toward DETA-NO compared to the wild type or corresponding complemented strains fhpA-egfp and gnoA-egfp. Deletion of both flavohemoglobin-encoding genes fhpA and fhpB further increased RNI sensitivity. When grown on AMM agar plates with NaNO₃ or NaNO₂ as nitrogen source, all the mutants lacking the gnoA gene revealed a slight growth defect (data not shown), most likely due to the delay in germination of the ΔgnoA mutant (Figure 2B). The delay in germination was reduced when ammonium tartrate or glutamine were used as sole nitrogen source. Determination of the dry weight revealed that growth of all mutants was reduced in comparison to the wild type when grown in liquid medium with nitrate as sole nitrogen source. Replacing nitrate with glutamine complemented the growth defect in all flavohemoglobin mutants, except in the ΔgnoA mutant strain (data not shown). These results were supported by northern blot analyses monitoring mRNA steady state transcript levels of fhpA and gnoA (Figure 3A). In AMM with nitrate as sole nitrogen source, the fhpA gene was strongly expressed, indicating the need to detoxify RNI accumulating during nitrate metabolism. When the fungus was cultivated with ammonium tartrate or glutamine, transcription of fhpA was comparatively low. In contrast to fhpA, transcript levels of gnoA were similar in all media tested and were not affected by the nitrogen source. Transcripts of fhpB were not detected under these conditions (data not shown).

To proof whether genes encoding RNI detoxifying enzymes are involved in the infection process, their transcription was analyzed in lungs of infected mice. For this purpose, total RNA extracted from lungs of infected mice was isolated and transcription of genes was analyzed by RT-PCR (Figure 3B). As positive control, transcription of the A. fumigatus act1 gene (actin; AFUA_6G04740) was monitored. The genes fhpA and gnoA were transcribed during the infection process, mRNA of fhpB was not detected.

To determine the localization of FhpA and GnoA in A. fumigatus, the respective genes were fused to the egfp gene under control of their native promoter. Transformation of the ΔfhpA or the ΔgnoA strain with the fhpA-egfp or gnoA-egfp construct, respectively, resulted in transformants displaying wild-type phenotype, as tested by resistance to DETA-NO in agar plate diffusion assays (Figures S7 and S8). This data indicated that the respective fusion proteins were functional. Fluorescence microscopy revealed localization of both the FhpA-eGFP (Figure 4A) and the GnoA-eGFP protein (Figure 4B) in the cytoplasm of all A. fumigatus morphotypes, from conidia to hyphae. Due to low transcriptional activity of the fhpB promoter, fluorescence of a strain producing FhpB-eGFP was not detectable (data not shown). Therefore, a mutant was generated that produces FhpB-eGFP under control of the constitutive otef promoter (Figure S9). Fluorescence microscopy of strain FhpB-eGFP verified localization of this flavohemoglobin in the mitochondria (Figure 4C).

To analyze whether the ability to detoxify macrophage-derived RNI influences fungal survival, the killing rate of wild type and mutant conidia confronted with macrophages was determined. No significant differences in killing of ΔfhpA, ΔfhpB, or ΔgnoA conidia by cells of the macrophage cell line MH-S were observed when compared to the wild type (Figure S10). This was also true for all double mutant strains deficient for RNI detoxification (data not shown).

The gnoA and fhpA mutant strains were tested in a mouse infection model for pulmonary aspergillosis. In this model, mice were immunosuppressed solely with cortisone acetate, i.e., neutrophils, that are able to produce NO, are still recruited to the site of infection. Survival of infected animals was monitored over a period of 14 days (Figure 5A). No significant differences in survival of mice infected with the mutant strains compared to mice infected with the wild type and the corresponding complemented strains occurred. Furthermore, histopathology of lungs of infected mice did not reveal any differences with regard to fungal growth or the recruitment of neutrophils (data not shown). In an alternative infection model, neutropenic mice, treated with cyclophosphamide, were infected with conidia of the ΔfhpA/ΔgnoA double deletion strain. Again, no difference in survival was detected compared to mice infected with wild-type conidia (Figure 5B). Additionally, virulence of the strain ΔfhpA/ΔfhpB was analyzed in a non-vertebrate model, using larvae of the greater wax moth G. mellonella. Also in this model, there was no difference in killing of larvae by the wild type or by the RNI detoxification mutants (Figure 5C).