Two C4-sterol methyl oxidases (Erg25) catalyse ergosterol intermediate demethylation and impact environmental stress adaptation in Aspergillus fumigatus

Abstract

The human pathogen Aspergillus fumigatus adapts to stress encountered in the mammalian host as part of its ability to cause disease. The transcription factor SrbA plays a significant role in this process by regulating genes involved in hypoxia and low-iron adaptation, antifungal drug responses and virulence. SrbA is a direct transcriptional regulator of genes encoding key enzymes in the ergosterol biosynthesis pathway, including erg25A and erg25B, and ΔsrbA accumulates C4-methyl sterols, suggesting a loss of Erg25 activity [C4-sterol methyl oxidase (SMO)]. Characterization of the two genes encoding SMOs in Aspergillus fumigatus revealed that both serve as functional C4-demethylases, with Erg25A serving in a primary role, as Δerg25A accumulates more C4-methyl sterol intermediates than Δerg25B. Single deletion of these SMOs revealed alterations in canonical ergosterol biosynthesis, indicating that ergosterol may be produced in an alternative fashion in the absence of SMO activity. A Δerg25A strain displayed moderate susceptibility to hypoxia and the endoplasmic reticulum stress-inducing agent DTT, but was not required for virulence in murine or insect models of invasive aspergillosis. Inducing expression of erg25A partially restored the hypoxia growth defect of ΔsrbA. These findings implicated Aspergillus fumigatus SMOs in the maintenance of canonical ergosterol biosynthesis and indicated an overall involvement in the fungal stress response.

Introduction

Aspergillus fumigatus is a filamentous fungus that can cause significant human disease in immunocompetent and immunocompromised hosts (Latgé, 1999; Pirofski & Casadevall, 2008; Tenholder, 1985). Medical treatments, such as anti-TNF therapy for autoimmune disorders, and the recent contamination of steroids used in joint therapy indicate that A. fumigatus is capable of causing disease in a myriad of patients and underlying immunocompromised states (Kainer et al., 2012; Pagano et al., 2006). Therefore, the study of A. fumigatus biology and virulence mechanisms is paramount to future patient health and therapeutic drug discovery.

Ergosterol biosynthesis is the pathway responsible for the construction of the fungal membrane sterol ergosterol, and is required for the stability and fluidity of the plasma membrane (Van Leeuwen et al., 2008; White et al., 2003). Along with ergosterol’s role as a target of antifungal drugs, these functions make the understanding of regulatory mechanisms for ergosterol biosynthesis of interest to biomedical research.

Importantly, ergosterol and some ergosterol biosynthesis intermediates are the targets of three of the four main antifungal drug classes: polyenes, allylamines and triazoles. Polyenes, including amphotericin B, target ergosterol and are thought to function by binding ergosterol in the membrane (Gray et al., 2012). Allylamines, including terbinafine, target the enzyme Erg1 (squalene epoxidase) and are used primarily for the treatment of dermatophytic infections (Petranyi et al., 1984). Triazoles target the enzyme Erg11A (also known as Cyp51A) (14α-demethylase) and may cause fungistatic activity by accumulation of toxic sterol intermediates (Kelly et al., 1995). Although these antifungal drugs represent our best treatment options for eradicating invasive Aspergillus infections, antifungal resistance and intrinsically azole-resistant species have been identified (Blum et al., 2013; Hadrich et al., 2012; Snelders et al., 2011a, b; Walsh et al., 2003). There are, however, other biochemical steps in the ergosterol biosynthesis pathway that are potential targets for antifungal drug development. Our ability to target these steps is currently unclear as less is known about the other steps of the ergosterol biosynthesis pathway and how the genes that encode these enzymes are regulated, especially in Aspergillus spp.

In A. fumigatus, the transcription factor SrbA is highly involved in transcriptional regulation of sterol biosynthesis (Willger et al., 2008). SrbA is a member of the sterol regulatory element binding protein (SREBP) helix–loop–helix family of transcription factors, and plays a significant role in fungal hypoxia and low-iron adaptation, triazole drug resistance and virulence (Blatzer et al., 2011; Blosser & Cramer, 2012; Walsh et al., 2003; Willger et al., 2008). The mechanism behind the severely attenuated hypoxia growth of the A. fumigatus SREBP-null mutant strain and its contribution to virulence remain unanswered questions. In human pathogenic fungi, hypoxia adaptation has been hypothesized to be a critical virulence attribute, as gene-deletion mutants that are unable to grow under hypoxia are typically also unable to cause invasive disease (Grahl et al., 2011; Willger et al., 2012). As a transcription factor, SrbA directly influences the expression of genes in its regulon, in which the ergosterol biosynthesis genes are prominent (Blatzer et al., 2011). Therefore, the phenotypes observed in the ΔsrbA strain may be attributed in part to the direct or indirect loss of specific ergosterol biosynthesis gene function.

Among SrbA’s regulon are the di-iron, di-oxygen-requiring enzymes Erg11 and Erg25 [C14-demethylase and C4-sterol methyl oxidase (SMO)]. This regulatory relationship is evidenced by the azole susceptibility and C4-methyl sterol intermediate accumulation observed in the ΔsrbA strain (Blosser & Cramer, 2012; Willger et al., 2008). Interestingly, removal of the C14-demethylase blockage by constitutive expression of erg11A in ΔsrbA aggravates the accumulation of C4-methyl sterols, suggesting that C4-demethylation is regulated by SrbA in A. fumigatus (Blosser & Cramer, 2012). This predominant C4-methyl sterol accumulation is also evidenced in the null mutant of the SREBP homologue in Cryptococcus neoformans, Sre1 (Chang et al., 2007; Lee et al., 2007). Furthermore, other critical phenotypes of ΔsrbA, such as hypoxia growth, are not remediated in the erg11A–ΔsrbA strain, indicating that other downstream targets of SrbA, such as SMOs, may be involved in these phenotypes.

Unlike Saccharomyces cerevisiae, the genome of A. fumigatus encodes two SMOs, erg25A and erg25B. Although the presence of two erg25-encoding genes is unusual in fungi, it is more commonplace in plants, where two SMOs each remove one of two C4-methyl sterols in a subsequent, but biochemically distinct mechanism (Benveniste, 2004; Darnet et al., 2001). In Arabidopsis thaliana, removal of each of the two C4-methyl groups requires a different SMO and mutants of the individual SMOs primarily accumulate their respective sterol intermediate.

Candida albicans also encodes two SMOs: erg25 and erg251. These SMOs do not have redundant function, however, as deletion of a single SMO is lethal and suggests that both copies work in tandem for efficient activity (Kennedy et al., 2000). These examples illustrate the diverse mechanisms that organisms have evolved to remove C4-methyl sterol intermediates in the eukaryotic kingdom.

In this study we examined the individual functions of the two SrbA-regulated SMOs in A. fumigatus. These paralogues each serve as a functional oxidase, as their respective genetic null mutants accumulated elevated levels of C4-methyl sterols. However, a probable difference in substrate affinity or timing of enzymic reaction is evidenced by Δerg25A accumulating more C4-methylated sterol intermediates than Δerg25B. These results suggested that Erg25A functions as the predominant or primary SMO in A. fumigatus, with Erg25B serving a secondary or subsequent enzymic function. Genetic elimination of both erg25 genes was suggested to be lethal due to the inability to generate such a strain, hinting that A. fumigatus compensated for the fragility of this critical juncture with the presence of two functioning SMOs.

Erg25A, as the primary SMO, appears to also play a role in adaptation to hypoxia and endoplasmic reticulum (ER) stress, but is not essential for virulence in a murine model of invasive aspergillosis. Importantly, by partially restoring erg25A mRNA abundance, we succeeded in reinitiating hypoxia growth in ΔsrbA. These data suggested that C4-methyl sterols were important for stress and hypoxia adaptation in A. fumigatus as part of the SrbA regulon. Regulation of ergosterol biosynthesis, and the key biochemical checkpoint at erg25, may be critical for signalling appropriate stress responses and marshalling a subsequent adaptive response in A. fumigatus.

Methods

Strains and media.

A. fumigatus strain CEA17 was used to construct the ΔsrbA, Δerg25A and Δerg25B strains (d’Enfert et al., 1999; Willger et al., 2008). To generate the erg25A (AFUB_084150) and erg25B (AFUB_098170) knockout constructs, ~1 kb upstream and downstream of both gene ORFs were PCR amplified from genomic A. fumigatus DNA. pyrG from Aspergillus parasiticus was amplified from plasmid pJW24 (courtesy of Dr Nancy Keller, University of Wisconsin Madison, Madison, WI, USA) and was inserted via fusion PCR (Yu et al., 2004). Protoplast-mediated transformation was conducted to introduce the foreign DNA into the CEA17 strain as described previously (Brown et al., 1998).

To reconstitute the WT erg25A or erg25B genes into their respective mutant background, a reconstitution construct was generated by cloning erg25A or erg25B from the WT strain CEA10 into fungal transformation vector pBC-Hygro (Dr Kevin McCluskey, Fungal Genetics Stock Center, Kansas City, MI, USA), which contained the hygromycin B resistance gene as a selectable marker.

The pflavA–erg25A : : ΔsrbA strain was constructed by PCR amplifying 1 kb upstream of the flavohaemoglobin ORF (AFUB_099650), the pyrG gene from A. parasiticus and the ORF of erg25A, then fusing these together via PCR. This pflavA–erg25A construct was transformed into the –pyrGΔsrbA background strain via protoplast-mediated transformation (Blosser & Cramer, 2012; d’Enfert et al., 1999).

Strains were verified via Southern blot for single, locus-specific integration, with the exception of the pflavA–erg25A : : ΔsrbA strain, which contained a single, random insertion. For the Δerg25A Southern blot, genomic DNA from the Δerg25A, WT and Δerg25A+erg25A strains was ApaI digested at 37 °C for 18 h. Strains were separated on a 1 % agarose gel, transferred and hybridized with a 607 bp probe. For the Δerg25B Southern blot, genomic DNA from the WT, Δerg25B and Δerg25B+erg25B strains was HindIII/NheI digested at 37 °C for 18 h. Strains were separated on a 1 % agarose gel, transferred and hybridized with a 328 bp probe. Copy insertion number was verified for pflavA–erg25A : : ΔsrbA via NcoI Southern blot with probes specific for srbA, pyrG and erg25A. The WT strain referred to in this paper is strain CBS 144.89 (i.e. CEA10).

All strains were grown routinely on glucose minimal media (GMM) that contained 1 % glucose, at 37 °C (Shimizu & Keller, 2001). The recipe for liquid GMM was identical to that for GMM, except without agar.

Hypoxia growth assay.

Growth under both normoxic (room air) and hypoxic (1 % O₂) conditions was evaluated utilizing a Ruskinn InvivO₂ 400 hypoxia chamber. Conidia (10⁴) of each strain were point inoculated in the centre of each agar plate. The radial growth at 37 °C under either normoxia or hypoxia was measured every 24 h for 3.5 days. Plates were photographed at 3.5 days. The experiment was repeated in biological triplicate.

Iron stress assay.

A. fumigatus (10⁸) conidia from various strains were cultured in 150 ml liquid GMM medium with or without iron (ferrous sulfate heptahydrate; Fisher Scientific) for 24 h. Following incubation, samples were harvested via vacuum filtration, frozen at −80 °C, lyophilized overnight and weighed.

ER stress assay.

A. fumigatus conidia (2×10⁷) were incubated in 35 ml liquid GMM at 37 °C with agitation, with or without DTT (1 mM; Sigma Aldrich) for 24 h. Cultures were harvested via vacuum filtration, frozen at −80 °C and lyophilized until dry. The biomass of the dried samples was measured.

Quantitative real-time (qRT)-PCR.

A. fumigatus strains were cultured in liquid GMM for 16 h in room air, then shifted to hypoxia for the indicated times. Mycelia were harvested via vacuum filtration and lyophilized overnight prior to homogenization with 0.1 mm glass beads. Total RNA was extracted using TRisure (Bioline) according to the manufacturer’s instruction and purified via the RNeasy column protocol (Qiagen). Genomic DNA elimination was completed with Turbo DNase I (Ambion). A secondary genomic DNA elimination was done with the Qiagen QuantiTect RT kit (Qiagen), as well as oligo-DT-primed cDNA synthesis. qRT-PCR was conducted in technical duplicates except where noted. A no-RT mRNA control was used to ensure removal of genomic DNA in each analysis.

Antifungal drug susceptibility testing.

Susceptibility to fluconazole, voriconazole and amphotericin B was evaluated with Etest strips (AB Biodisk). Strips were placed on an RPMI-MOPS (Sigma Aldrich) agar plate, pH 7.0, containing a lawn of 10⁵ conidia. Growth inhibition was evaluated at 48 h by visualization of growth inhibition when grown at 37 °C.

Calcofluor white cell wall perturbation assay.

GMM plates containing 25 µg ml⁻¹ calcofluor white were spotted with 5 µl of a 0.05 % Tween 80 solution containing 10⁶ conidia. Plates were incubated at 37 °C and radial growth measured every 24 h for a total of 96 h.

Total ergosterol assay.

Total ergosterol from A. fumigatus strains was extracted from mycelia after 24 h shaking growth in liquid GMM in room air. Mycelia were harvested via vacuum filtration, dried and weighed, and 200 mg dry fungal tissue extracted as described previously (Alcazar-Fuoli et al., 2008; Blosser & Cramer, 2012). Total ergosterol was measured using a Shimadzu CLASS-VP HPLC and detected via a SPD-M10AVP PDA at 280 nm on a Bondapak C18 column (300 mm×3.9 mm, 10 µm). The quantity of ergosterol per strain was calculated from a standard curve of ergosterol (Acros Organics).

Sterol profile.

Mycelia were grown and harvested as described for total ergosterol extraction. Extraction was done as described previously (Alcazar-Fuoli et al., 2008; Blosser & Cramer, 2012) and was identical to in the total ergosterol extraction with the exception that the neutral lipids were extracted with 1.5 ml hexane. Vials were stored at −20 °C until ready for injection. Samples were resolved in 100 µl toluene and derivatized with 100 µl BFSTA (N,O-bis(trimetylsilyl)trifluoroacetamid-trimethylchlorosilane) for 1 h at 70 °C. Sterols were measured via GC-MS, using a Shimadzu QP-2010 GC with a quadrupole mass ionizer. Peaks were identified via electron ionization fragmentation pattern and retention time (Table S1, available in the online Supplementary Material). All data were relative to total ergosterol content. The GC program and MS operation conditions were similar to those described previously (Alcazar-Fuoli et al., 2008; Blosser & Cramer, 2012), with the following exceptions: the oven was maintained at 180 °C for 2 min, then heated to 310 °C at 10 °C min⁻¹ and finally held at 310 °C for 15 min.

Virulence assays.

Galleria mellonella larvae were used in an insect model of invasive aspergillosis (Bergin et al., 2005; Reeves et al., 2004). Twenty larvae per strain were injected, with 10⁵ conidia in 10 µl Tween 80 (0.05 % solution), into the larval haemocoel through the last pro-leg using a Hamilton syringe. A Tween injection control was included. Larvae were maintained in a wood-shaving-filled Petri dish, incubated at 37 °C and monitored for mortality daily. This experiment was independently performed three times; analysis represents the pooled survival data from the three experiments.

CD1 female mice, 6–8 weeks old, were used in a corticosteroid model of invasive pulmonary aspergillosis. Mice were obtained from Charles River Laboratories. Five mice per cage were housed in a HEPA-filtered environment with autoclaved food ad libitum. For the corticosteroid model, a single 40 mg kg⁻¹ injection of Kenalog (Bristol-Myers Squibb) was administered subcutaneously 1 day prior to inoculation with A. fumigatus. For inoculation, 2×10⁶ conidia in 100 µl 10 % PBS were administered intranasally to 10 mice per group under isoflurane anaesthesia. Mice were observed for 14 days after A. fumigatus challenge. Any animal showing severe distress (defined as ruffling of fur, hunched posture, decreased mobility and difficulty in breathing) was euthanized humanely and recorded as deceased within that 24 h time period. A PBS-only control group was used to verify the underlying health status of these animals. No PBS-only animals perished during the course of the experiment.

Ethics statement.

This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The animal experimental protocol was approved by the Institutional Animal Care and Use Committee at Montana State University (Federal-Wide Assurance Number A3637-01).

Statistical analysis.

All results presented with statistical significance were analysed with an unpaired two-tailed Student’s t-test, with the exception of the virulence assay. Results from the insect and murine virulence assay were calculated via log-rank (Mantel–Cox) analysis of the Kaplan–Meir survival curve.

Results

SMO identification and homology in A. fumigatus

Erg25 (SMO) is a non-haem iron-requiring enzyme, characterized by three histidine motifs that are conserved throughout Eukarya (Fig. 1) (Bard et al., 1996; Darnet et al., 2001; Li & Kaplan, 1996; Shanklin et al., 1994). These three histidine motifs, HX₃H, HX₂HH and HX₂HH, are characteristic of the sterol desaturase-like family, which includes the C5-sterol desaturases (Erg3) and the SMOs (Erg25) (Shanklin et al., 1994). These motifs are postulated to be iron-binding sites and critical for enzymic function. Although these histidine motifs are highly conserved in the SMOs of S. cerevisiae (ScErg25), Homo sapiens (MSMO1), C. albicans (CaErg25 and CaErg251), and A. fumigatus (Erg25A and Erg25B), the divergence of SMO inter- and intra-species amino acid sequence is significant (Bard et al., 1996; Kennedy et al., 2000; Li & Kaplan, 1996).

Fig. 1.

Comparative amino acid homology and motif analysis of Aspergillus fumigatus Erg25 with previously characterized SMOs throughout Eukarya. Dark-grey-shaded boxes indicate conserved histidine motifs in Saccharomyces cerevisiae (Sc), Homo sapiens (Hs), Candida albicans (Ca) and A. fumigatus (Af). Shaded unboxed residues are conserved in at least one homologue in S. cerevisiae, C. albicans and A. fumigatus.

Erg25A and Erg25B share 53 % amino acid identity with each other. Interestingly, amino acid alignment with other fungi, such as Neosartorya fischeri, Coccidioides immitis, Neurospora crassa, Schizosaccharomyces pombe, S. cerevisiae and C. albicans, indicates that fungal SMOs appear to be more closely related to Erg25B than Erg25A (60.2±19.8 versus 46.8±6.7 % amino acid homology, respectively). This analysis suggested that A. fumigatus Erg25A is evolutionarily more distant from its SMO fungal homologues than Erg25B.

Construction of SMO-null mutants: Δerg25A and Δerg25B

Classically, genetic deletions of SMOs are lethal or, at a minimum, highly deleterious to the mutated organism (Kennedy et al., 2000; Li & Kaplan, 1996). Many of the genes of A. fumigatus have a paralogue, owing to gene duplication (Nierman et al., 2005), and duplication of ergosterol biosynthesis genes has been hypothesized to be an adaptation for membrane composition and integrity (da Silva Ferreira et al., 2006; Nierman et al., 2005). As A. fumigatus has retained classical SMO protein architecture in both erg25A and erg25B, we hypothesized that single genetic replacement of either gene would not be lethal, but instead would provide functional compensation.

To determine the role of individual SMOs in A. fumigatus, single genetic replacement mutants were constructed. These strains were viable, and were validated for single, locus-specific insertion via PCR and Southern blot analysis (Fig. S1). These mutants demonstrated no difference in sporulation, or growth on liquid or solid minimal medium compared with the WT strain under normoxic laboratory conditions (Fig. S1 and data not shown). Therefore, A. fumigatus SMOs appear to be able to compensate for loss of their paralogue because single deletion of either gene was not lethal, as observed in other organisms (Kennedy et al., 2000; Li & Kaplan, 1996).

Two viable SMOs: sterol intermediate profile verification

A. fumigatus SMOs appear to be functional oxidases, as Δerg25A and Δerg25B sterol profiles demonstrated accumulation of C4-methyl sterols, stereotypical for blockage or elimination of SMO function. In A. fumigatus, the C4-methyl sterol intermediates were 4-methyl fecosterol and 4,4-dimethyl fecosterol (Figs 2a and S2). The C4-methyl sterol build-up was especially notable for Δerg25A, which accumulated 17- and 32-fold more 4-methyl fecosterol and 4,4-dimethyl fecosterol, respectively, compared with WT (Fig. 2a; P<0.01). Δerg25B, conversely, only accumulated 1.5- and 7-fold more 4-methyl fecosterol and 4,4-dimethyl fecosterol compared with WT levels, which was not statistically significant (Fig. 2a). These data suggest that Erg25A serves as the predominant SMO in A. fumigatus or may be the first step of two sequentially acting SMOs. More mechanistic studies would be required to elucidate the precise function of these enzymes, but this step-wise progression of enzymic action has been demonstrated using virus-induced gene silencing of SMOs in Nicotiana benthamiana (Darnet & Rahier, 2004).

Fig. 2.

Characterization of SMO mutants. (a) Significant accumulation of 4-methyl fecosterol [4,24-dimethyl-cholesta-8,24(28)-dien-3β-ol] and 4,4-dimethyl fecosterol [4,4,24-trimethyl-cholesta-8,24(28)-dien-3β-ol] was observed in ΔsrbA and Δerg25A when compared with the WT strain. Increased accumulation of these intermediates was also observed in Δerg25B, but was not statistically significant. Results are displayed as the ratio of sterol intermediates to ergosterol. Peaks were identified as described previously (Blosser & Cramer, 2012) and in Methods. *P<0.05. (b) Total ergosterol is unaltered in Δerg25A and Δerg25B compared with the WT strain, in contrast to the observed depletion of total ergosterol in ΔsrbA. **P<0.0001. Values represent the mean+sd of three or four biological and three technical replicates for total ergosterol, and two or three biological replicates for sterol intermediate profiles.

ΔsrbA hyper-accumulates both C4-methyl sterol intermediates (Fig. 2a; P<0.01) (Blosser & Cramer, 2012, Willger et al., 2008) and displays a significant (30–70 %) decrease in total ergosterol (Fig. 2b; P<0.01) (Blosser & Cramer, 2012). Although C4-methyl sterol intermediates disproportionally accumulated, particularly in Δerg25A, total ergosterol was not significantly altered in either erg25 genetic null mutant (Fig. 2b). Therefore some, but not all, ergosterol biosynthesis defects observed in ΔsrbA may be attributed to diminished erg25A and/or erg25B activity.

SMO-null mutant sterol profiles reveal altered C9(11)-intermediate accumulation

Alterations in sterol intermediate profiles illuminate the regulatory and compensatory mechanisms an organism utilizes to produce ergosterol under suboptimal conditions (Alcazar-Fuoli et al., 2008; Darnet et al., 2001; Ott et al., 2005). Δerg25A and Δerg25B tended to accumulate the sterol intermediates 9(11)-dehydroergosterol and parkeol (Figs 3a and S2). Parkeol was especially evident in Δerg25B, whereas 9(11)-dehydroergosterol accumulated at elevated levels in both Δerg25A and Δerg25B (Fig. 3a). 9(11)-Dehydroergosterol also trended toward accumulation in ΔsrbA. These findings may suggest that an alternative conversion mechanism exists for traversion from squalene to ergosterol in A. fumigatus. In particular, the accumulation of parkeol is contraindicative of canonical ergosterol biosynthesis, as the 9(11) bond of parkeol is not thought to be conducive to C14- or C4-demethyalation (Venkatramesh & Nes, 1995). These findings support a non-linear model of ergosterol biosynthesis, as suggested previously by Alcazar-Fuoli et al. (2008).

Fig. 3.

Sterol intermediate profiles for Δerg25A and Δerg25B, and proposed alternative pathway. (a) Accumulation of 9(11) intermediates in Δerg25A, Δerg25B and ΔsrbA. Much higher parkeol [4,4,14-trimethyl-cholesta-9(11),24-dien-3β-ol] is observed in Δerg25B than in the other strains studied. Loss of erg25A or erg25B resulted in accumulation of 9(11)-dehydroergosterol [24-methyl cholesta-5,7,9(11),22-tetraen-3β-ol]. Results represent the ratio of sterol intermediate to ergosterol. (b) Schematic of ergosterol biosynthesis pathway (left) and proposed alternative mechanism of parkeol and 9(11)-dehydroergosterol intermediate accumulation (right). Dotted arrows indicate the proposed mechanism; solid arrows indicate the mechanism proposed by Alcazar-Fuoli et al. (2008) (c) Aberrant accumulation of sterol intermediates involving passage from episterol to ergosterol. 22-Dihydroergosterol [24-methyl cholesta-5,7-dien-3β-ol], ergostatetraenol [24-methyl cholesta-5,7,22,24(28)-tetraen-3β-ol] and 5-dihydroergosterol [24-methyl cholesta-5,7-dien-3β-ol]. *P<0.05; **P<0.001; ns, not significant. Analysis displayed was between the WT and indicated mutant strain, with the exception of 22-dihydroergosterol, which was analysed by comparing Δerg25B to ΔsrbA and Δerg25A, due to variability observed in the WT strain. No difference in 22-dihydroergosterol accumulation was displayed when WT was compared to ΔsrbA, Δerg25A or Δerg25B.

Strains that accumulated parkeol also trended toward accumulation of 9(11)-dehydroergosterol (Fig. 3a), suggesting either that there are novel mechanisms for C14-demethylation of parkeol, or that C8(9) or C9(11) bonds can isomerize in A. fumigatus (Figs 3b and S2) (Venkatramesh & Nes, 1995). Alternatively, 9(11)- dehydroergosterol could also represent increased oxidation of ergosterol in these mutants. It is therefore plausible that canonical sterol biosynthesis is altered due to accumulation of aberrant sterol intermediates in SMO mutants (Fig. 3b). The most striking evidence of this effect was the observation that the described intermediates did not accumulate at elevated levels in WT strains (Alcazar-Fuoli et al., 2008; S. J. Blosser & R. A. Cramer, unpublished observation), but only when severe environmental perturbation or genetic mutation was encountered (Blosser & Cramer, 2012; S. J. Blosser & R. A. Cramer, unpublished observation).

Alternative mechanism for episterol-to-ergosterol transition

Although ergosterol biosynthesis is linear for much of the pathway, the steps between episterol and ergosterol are branched, allowing for three step-wise uses of the enzymes Erg5, Erg4 and Erg3. It has been postulated that the order of action of these enzymes is unimportant for the overall process of ergosterol biosynthesis (Alcazar-Fuoli et al., 2008; da Silva Ferreira et al., 2005).

SMO-null mutants, however, displayed significant alterations in how they transversed between episterol and ergosterol. In Δerg25A, only the Erg3→Erg4→Erg5 mechanism appeared to be fully functional, as ergostatetraenol was not detectable in this strain and 22-dihydroergosterol accumulated at slightly aberrant levels when compared with the WT strain (Figs 3c and S2). 22-Dihydroergosterol appeared to accumulate stably in both Δerg25A and Δerg25B, with abundant accumulation observed in Δerg25B. This intermediate was not always detected in WT strains (0.83 % accumulation in replicate one, zero accumulation in replicates two and three), suggesting that both erg25A and erg25B are required for optimal removal of this intermediate. The data may indicate that the Erg3/Erg4→Erg5 transition was not the preferred transition for A. fumigatus strains containing both erg25A and erg25B.

Interestingly, 5-dihydroergosterol and 22-dihydroergosterol, present in both Δerg25A and Δerg25B, were absent in ΔsrbA (Fig. 3c). This is intriguing given the regulation by SrbA of not only erg25A/B, but also erg3B and erg5 (Losada et al., 2014). We observed a different pattern with the sterol intermediate ergostatetraenol, however, as ΔsrbA produced diminished quantities of this intermediate, Δerg25A produced no detectable intermediate and Δerg25B had a normal abundance of ergostatetraenol. These results suggested that only Erg25B could catalyse this transition (Fig. 3c). Therefore, it is possible that either Erg25A or Erg25B function allowed for the mechanism Erg5→Erg4→Erg3 to occur, whereas other mechanisms (Erg3/Erg5→Erg4 and Erg3/Erg4→Erg5) were more specific for Erg25A and Erg25B, respectively.

Alternatively, the presence of Erg25A or Erg25B may have altered the stability/function of Erg3, Erg4 or Erg5 to interact with downstream sterol intermediates as part of the ergosome (Mo & Bard, 2005; Mo et al., 2004). Taken together, these data suggested that Erg25A and Erg25B have complementary, but not identical functions, particularly as demonstrated by the differences in sterol intermediate accumulation in Δerg25A and Δerg25B.

SMO genes responsive to hypoxia

The transcription factor SrbA binds directly to the promoters of both erg25A and erg25B (Blatzer et al., 2011; Losada et al., 2014). As SrbA is essential for hypoxia adaptation, we examined the response of erg25A and erg25B to hypoxia exposure. srbA mRNA was increased significantly after 30 min and 2 h exposure to 1 % oxygen compared with normoxia (Fig. 4a; **P<0.0001). mRNA levels of both SMOs were also increased significantly upon exposure to hypoxia, with 13- and 3-fold induction of erg25A and erg25B upon 30 min incubation with hypoxia, and 20- and 7-fold induction after 2 h (Fig. 4a; *P<0.0001, hypoxia versus normoxia). Given the similar accumulation of C4-methyl sterol intermediates in ΔsrbA and Δerg25A, it is perhaps surprising that radial growth rates of Δerg25A (or Δerg25B) were not affected under hypoxia (1 % O₂) when compared with growth under normoxia (Fig. 4b). A further reduction in oxygen (0.2 % O₂) level, however, elicited the copious production of aerial hyphae in Δerg25A, indicative of an increased stress response to hypoxia (Fig. 4b) (McCormick et al., 2012; Tamame et al., 1983). Taken together, these results suggested that loss of hypoxic growth in ΔsrbA is not attributed exclusively to accumulation of C4-methyl sterol intermediates.

Fig. 4.

SMOs are responsive to hypoxia. (a) qRT-PCR analysis of erg25A, erg25B and srbA in response to 30 min and 2 h exposure to hypoxia. WT Aspergillus fumigatus was grown for 16 h at 37 °C and then harvested (N 0 h) or exposed to hypoxia (1 % O₂) for 30 min (H 30 min) or 2 h (H 2 h) prior to harvesting. RNA was extracted and cDNA synthesized. The normalized fold expression represents the mean and percentage error of three biological replicates normalized to the housekeeping gene tefA and was relative to the N 0 h sample for all genes examined. Data represent three biological and two combined technical replicates. Analysis compared mRNA transcript abundance in mutant strains to the WT. **P<0.0001. (b) Radial growth assay of SMO strains under hypoxia (1 % O₂) and extreme hypoxia (0.2 % O₂) revealed a stress-response phenotype in Δerg25A but not in Δerg25B. Conidia (10⁵) were spotted on a GMM plate and allowed to grow at 37 °C for 96 h under the conditions indicated. (c) Conidia (10⁴) were spotted on solid media (GMM) and allowed to grow at 37 °C for 4 h under normoxia (17 % O₂) or hypoxia (1 % O₂). Plates were representative of three biological replicates per strain under each condition. (d) qRT-PCR of pflavA–erg25A : : ΔsrbA revealed an increase in erg25A mRNA abundance relative to levels in ΔsrbA; however, srbA transcript levels remained unchanged. Conidia for WT, ΔsrbA or pflavA–erg25A : : ΔsrbA were grown for 16 h at 37 °C and then transferred to hypoxia for 2 h. pflavA–erg25A–ΔsrbA levels of erg25A were significantly higher than in ΔsrbA. Data represent three biological and two technical replicates. mRNA abundance was normalized to the housekeeping gene tefA and is relative to the WT strain. **P<0.0001.

In further support of this hypothesis, we uncoupled erg25A’s dependency on SrbA for transcriptional activation by placing erg25A under a strong nitrogen-regulated promoter (flavohaemoglobin gene flavA; AFUA_4G03410/AFUB_099650) within the ΔsrbA background. Experimental evidence suggested that AFUB_099650 mRNA increased dramatically when exposed to hypoxia when using nitrate as its sole nitrogen source (Losada et al., 2014; Schinko et al., 2010; Zhou et al., 2009). pflavA–erg25A : : ΔsrbA was verified to have a single insertion by PCR and Southern blot analysis (Fig. S3). pflavA–erg25A : : ΔsrbA significantly increased erg25A mRNA abundance in ΔsrbA, but did not fully restore mRNA levels to WT levels (Fig. 4c; **P<0.001). Consistent with Erg25 playing a critical role in adaptation to stress, hypoxic growth was restored partially to ΔsrbA when erg25A mRNA levels were increased in ΔsrbA (Fig. 4d).

Aberrant stress response in Δerg25A and Δerg25B

Alterations in erg25 mRNA abundance in fungi have previously been characterized during stress-inducing conditions in C. neoformans, C. albicans, S. cerevisiae and Trichophyton rubrum (Bammert & Fostel, 2000; Barker et al., 2004; Borecká-Melkusová et al., 2009; De Backer et al., 2001). Intriguingly, erg25 mRNA abundance appeared to be highly elevated in stress-inducing conditions of highly diverse origin, including triazole antifungal drugs, hypoxia, dysregulated calcium signalling and lipid homeostasis, when encountering ER or oxidative stress, and during biofilm formation (Bammert & Fostel, 2000; Borecká-Melkusová et al., 2009; da Silva Ferreira et al., 2006; Diao et al., 2009; Feng et al., 2011; Florio et al., 2012; Hameed et al., 2011; Hughes et al., 2007; Lee et al., 2007; Nailis et al., 2010). As the aforementioned stress-inducing stimuli were multifaceted in their mechanisms, we hypothesized that Erg25 function was critical for ubiquitous stress adaptation, possibly as an integral portion of the ‘fungal adaptome’ (Hartmann et al., 2011).

To test this hypothesis, we subjected Δerg25A and Δerg25B to potential cellular stresses encountered by A. fumigatus in a mammalian host. Overall, the effect of SMO loss had modest effects on the stress response of A. fumigatus, possibly due to the ability of one paralogue to compensate for the loss of the other. However, observed differences between the two mutant strains suggested that sterol intermediates may influence the stress response in A. fumigatus in a diverse fashion, depending primarily on the nature of the response.

ER stress.

We examined the response of the SMO-null mutants to 1 mM DTT, which interferes with disulfide bond formation and ultimately causes protein unfolding (Feng et al., 2011). ΔsrbA, Δerg25A and Δerg25B all displayed significant susceptibility to DTT when compared with the WT (Fig. 5a; **P<0.0001). Growth of both Δerg25A and ΔsrbA was inhibited similarly by DTT, with growth ratios 50.5 and 46.5 % less (untreated versus treated) than the WT. The Δerg25B strain was more susceptible to DTT, with a 62.9 % reduction in growth ratio compared with the WT. Reconstitution of srbA, erg25A or erg25B fully restored tolerance to DTT (Fig. 5a). Overall, loss of SMO function conferred susceptibility to DTT, demonstrating the requirement for SMO function in either membrane stability or other potential signalling pathways in A. fumigatus.

Fig. 5.

Aberrant stress responses in SMO mutants. (a) DTT ER stress-inducing assay. Conidia (2×10⁷) were incubated in liquid GMM with or without DTT (1 mM) for 24 h. Ratios express the biomass of culture with DTT treatment versus control. Data are the mean±sd of three biological replicates per strain. Increasing susceptibility was observed for ΔsrbA, Δerg25A and Δerg25B compared with WT, **P<0.0001. In addition, Δerg25B was significantly more susceptible than ΔsrbA and Δerg25A, P<0.05. Restoration of srbA, erg25A or erg25B into the corresponding mutant background restored observed DTT susceptibility. (b) Iron biomass assay. Aspergillus fumigatus conidia (10⁸) from various strains were cultured in 150 ml liquid GMM with or without iron (ferrous sulfate heptahydrate) for 24 h. Data represent the ratio of growth without iron versus with iron. The mean±sd of three or more biological replicates is reported. In contrast to the increased susceptibility of ΔsrbA to low iron, both Δerg25A and Δerg25B were significantly more adapted to growth in low iron than the WT strain. Reconstitution of WT srbA, erg25A or erg25B in the respective strain ameliorated any observed alterations in iron stress adaptation. *P<0.05; **P<0.0001; ns, not significant. (c) Antifungal MIC determination using Etests. FLC, fluconazole; AMB, amphotericin B; VCZ, voriconazole. Ranges represent variation between biological replicates. (d) Enhanced resistance to calcofluor white observed in Δerg25A and Δerg25B when compared with WT. Agar plates containing 25 mg/ml of calcofluor white were inoculated with 10⁶ conidia suspended in a 0.05 % solution of Tween 80. Plates were incubated at 37C and radial growth was measured every 24 h. Data are represented as the mean±sd radial growth (cm) of the individual colonies, and as percentage growth diameter increase when compared to WT of three biological replicates per strain.

Iron stress.

Because Erg25 is a di-iron-requiring enzyme and thus integrally impacted by the cellular iron status, we subjected Δerg25A and Δerg25B conidia to iron-depleted conditions and evaluated subsequent biomass. Briefly, WT, mutant strains and reconstituted strains were cultured in iron-replete (30 µM) or iron-free media for 24 h. Significantly, both Δerg25A and Δerg25B exhibited increased growth in low-iron conditions compared with the WT strain (iron depleted versus iron replete; Fig. 5b) and their respective reconstituted strains. These findings were inverse to what was observed with ΔsrbA, which demonstrated a significant growth reduction under low-iron environments (Fig. 5b) (Blatzer et al., 2011).

Triazole antifungal drugs.

Treatment of C. albicans, C. neoformans, A. fumigatus or T. rubrum with triazole antifungal drugs elicited a robust increase in erg25 mRNA abundance (Bammert & Fostel, 2000; Barker et al., 2004; da Silva Ferreira et al., 2006; Diao et al., 2009). There is some suggestion that 4-methyl sterol intermediates may provide additional layers of control for ergosterol biosynthesis, as treatment with triazole antifungal drugs ubiquitously results in an increase in erg25 mRNA abundance. To determine the role of erg25 in mediating resistance to triazole antifungal drugs in A. fumigatus, we subjected Δerg25A and Δerg25B to fluconazole and voriconazole, and to the ergosterol-targeting polyene, amphotericin B. Neither Δerg25A nor Δerg25B exhibited altered susceptibility to the antifungal drugs tested (Fig. 5c). This was in direct contrast to ΔsrbA that showed significant susceptibility to both fluconazole and voriconazole, which has been linked to a decrease in erg11A mRNA transcript levels (Blosser & Cramer, 2012). Interestingly, ScΔerg25 demonstrated altered mitochondrial physiology and appearance, which has been linked to altered toxicity of triazole antifungal drugs (Altmann & Westermann, 2005; Kontoyiannis, 2000). Additionally, the SMO-null mutants did not display increased susceptibility to the cell wall-perturbing agent Congo red, high-temperature stress (50 °C) or H₂O₂ (data not shown). A modest increase in resistance to calcofluor white and a modest increase in susceptibility to paraquat was observed in Δerg25A (Fig. 5d and data not shown). Taken together, these data suggest that loss of SMOs moderately impacted the ability of A. fumigatus to tolerate various stress-inducing conditions.

Erg25A is not required for virulence of A. fumigatus

SMOs are essential genes in S. cerevisiae and C. albicans (Gachotte et al., 1997; Kennedy et al., 2000). As the two A. fumigatus SMOs appear to function largely independently of one another, we had the unique opportunity to assess individual function and contribution to virulence in the fungal kingdom with this system. To assess the contribution of single SMOs to virulence, we assessed Δerg25A and Δerg25B in a G. mellonella model of invasive aspergillosis (Bergin et al., 2005; Reeves et al., 2004). No change in virulence was observed in this insect model with either Δerg25A or Δerg25B compared with the WT strain (Fig. 6a).

Fig. 6.

Erg25A was not required for virulence of Aspergillus fumigatus. (a) Conidia (10⁵) were injected into the haemocoel of Galleria mellonella larvae in an insect model of invasive aspergillosis. Survival of larvae was assessed daily for 5 days post-inoculation. A Tween-only (Tween 80, 0.05 % solution) control was included as an injection control. Data represent the combined results of three independent experiments. No virulence defect was observed in any of the strains examined. (b) Conidia (2×10⁶) were administered intranasally to corticosteroid-treated (Kenalog, 40 mg kg^–1) CD1 mice from WT, PBS-only and Δerg25A strains. Survival was assessed for 14 days post-inoculation. The SMO mutant Δerg25A did not show any alteration in virulence compared with the WT strain.

A. fumigatus mutants susceptible to ER stress-inducing agents and hypoxia may have altered virulence in murine models of invasive aspergillosis that are not apparent in the G. mellonella model due to differing host parameters (Feng et al., 2011; Hartmann et al., 2011; Rahier, 2011; Snelders et al., 2011a). As Δerg25A demonstrated increased susceptibility to both DTT and hypoxia, we examined whether erg25A function was critical for the virulence of A. fumigatus using the corticosteroid model of invasive pulmonary aspergillosis, which is induced by the synthetic steroid triamcinolone. The corticosteroid model is known for its abundant inflammation due to robust immune cell recruitment, which contributed to the significant incidence of hypoxia in this model (Grahl et al., 2011). Elimination of erg25A alone was not sufficient to alter the virulence of A. fumigatus in this model when compared with a WT strain (Fig. 6b).

Discussion

In fungi, the maintenance of ergosterol biosynthesis is a critical component of fungal growth, but its overall impact on fungal biology, particularly in the context of virulence, remains to be fully elucidated. The ability of A. fumigatus to survive in vivo requires several key components, termed the ‘fungal adaptome’ (Hartmann et al., 2011). These attributes, amongst others, include the ability to grow in a low-oxygen, low-iron and reactive-oxygen-species-rich niche, provide some of the arsenal required to germinate, penetrate tissue and cause disease. Fungal members of the SREBP transcription factor family, including A. fumigatus SrbA, have been shown to be critical for fungal virulence (Chang et al., 2007; Lee et al., 2007; Willger et al., 2008). A major component of the fungal SREBP transcriptional regulons in both C. neoformans and A. fumigatus is ergosterol biosynthesis. We observed occupancy of SrbA at the erg25A putative promoter region in addition to a significant decrease in both erg25A and erg25B mRNA levels in a srbA genetic null mutant (Blatzer et al. 2011, Willger et al., 2008). Moreover, ΔsrbA accumulates significant levels of C4-methyl sterols (Blosser & Cramer, 2012). We therefore set out to further understand how erg25A and erg25B not only assist A. fumigatus in adapting to the conditions encountered in the mammalian host, but also how the reduction of erg25A and erg25B in ΔsrbA contributes to its significant phenotypes.

Ergosterol biosynthesis requires 12 molecules of oxygen to progress from squalene to ergosterol (Chang et al., 2007; Lee et al., 2007). In A. fumigatus, six reactions involving at least 16 different enzymes are employed, four of which require iron as a cofactor and six of which require NADPH (Alcazar-Fuoli et al., 2008; Chang et al., 2007). C4-demethylation, one of the iron-requiring steps, catalysed by Erg25, occurs twice in the squalene-to-ergosterol transition and consumes six of the 12 requisite oxygen molecules. Therefore, it is quite plausible that C4-demethylation is a rate-limiting step. This hypothesis is bolstered by the robust connection between oxygen and iron requirements in ergosterol biosynthesis, and the regulation of ergosterol biosynthesis and iron acquisition by a key hypoxia-coordinated transcription factor, SrbA (Blatzer et al., 2011).

A. fumigatus has maintained two C4-demethylase genes, which encode two distinct demethylases, Erg25A and Erg25B (Fig. 1). Although not unique in the fungal kingdom, prior attempts to generate null mutants of other fungal SMOs have largely been unsuccessful, with rescue achieved only by blocking C14-demethylation with triazole antifungal drugs (Gachotte et al., 1997). In C. albicans, which has also maintained two paralogues of erg25, deletion of CaErg25 results in a strain that is not viable. These findings suggest that CaErg251 is either a non-functioning SMO or that CaErg251 cannot demethylate adequately without CaErg25 (Kennedy et al., 2000). Interestingly, single SMO mutants did not cause lethality in A. fumigatus. However, both Δerg25A and, to a lesser extent, Δerg25B accumulated C4-methyl sterols, indicating they function as SMOs in A. fumigatus (Fig. 2a). Genetic replacement of either SMO, however, did not elicit any of the other predominant erg25 mutant phenotypes characterized in the literature. These genetic and phenotypic data suggest that Erg25A and Erg25B in A. fumigatus may have partial functional redundancy. Further support for this hypothesis comes from multiple unsuccessful attempts to generate a complete Erg25 genetic null mutant by gene replacement of both erg25A and erg25B (data not shown). However, the large accumulation of C4-methyl sterol intermediates in Δerg25A and the modest, but apparent, stress-related phenotypes observed in Δerg25A, but largely not Δerg25B, suggest that Erg25A and Erg25B have evolved distinct functions in A. fumigatus.

Although the purpose of maintaining two SMOs within A. fumigatus remains unknown, it is plausible that substrate availability may determine enzymic usage for C4-demethylation or simply that Erg25B serves in an auxiliary function to Erg25A. Importantly, maintenance of two SMOs with similar substrate specificity has not been characterized previously in Eukarya. Perhaps the closest model for our data is found in the well-studied plant A. thaliana. In A. thaliana, two copies of erg25 also exist and both serve as functional SMOs (Darnet & Rahier, 2004); deletion of either SMO is not lethal and single mutants accumulate C4-methylated sterol intermediates, much as in A. fumigatus (Darnet & Rahier, 2004). The entities accumulated, however, are not identical in both mutants, as ΔSMO1 accumulates mainly 4,4-dimethyl-9β,19-cyclopropylsterols and ΔSMO2 accumulates mainly 4-methyl-Δ⁷-sterols (Darnet & Rahier, 2004). These plant-specific sterol intermediates correspond roughly to an accumulation in 4,4-dimethyl fecosterol and 4-methyl fecosterol, respectively, in fungi. Akin to what is observed in A. fumigatus, it appears that A. thaliana encodes distinct sets of SMOs that are more substrate specific than other eukaryotes studied to date. It appears that in A. fumigatus, Erg25A functions as the predominant SMO given Δerg25A’s pronounced accumulation of C4-methyl sterol intermediates compared with WT and Δerg25B (Fig. 2a). This hypothesis is validated by the other visible phenotypes of Δerg25A that are not apparent in Δerg25B, including changes in 22-dihydroergosterol and ergostatetraenol accumulation (Fig. 3c), high mRNA induction in hypoxia (Fig. 4a), enhanced stress response in extreme hypoxia (Fig. 4b), and susceptibility to DTT (Fig. 5a). It is possible that Erg25B functions as a back-up for Erg25A in its C4-demethylation function or may supply structural efficiency for Erg25A. This latter suggestion is based on the concept of an ergosterol ‘hub’ or complex (i.e. ergosome), the maintenance of which is required for full enzymic efficiency of ergosterol biosynthesis (Mo & Bard, 2005; Mo et al., 2004). Thus, in addition to their direct enzymic function, these enzymes may have critical structural roles that affect the function of an enzyme complex that could impact their biological functions.

A review of the current Erg25 or SMO literature reveals the biological importance of these enzymes in response to various cellular stressors. SMO mRNA is highly induced during stress-inducing conditions in many other fungi (Bammert & Fostel, 2000; Barker et al., 2004; Borecká-Melkusová et al., 2009; De Backer et al., 2001), including antifungal drug treatment, during hypoxia and during infection (Bammert & Fostel, 2000; Borecká-Melkusová et al., 2009; da Silva Ferreira et al., 2006; Diao et al., 2009; Feng et al., 2011; Florio et al., 2012; Hameed et al., 2011; Hughes et al., 2007; Lee et al., 2007; Nailis et al., 2010). This interplay between hypoxia growth, ergosterol biosynthesis and stress has been a common thread in our studies, and our results suggest that these interactions are controlled tightly by central regulatory components.

Growth under hypoxia results in the enrichment of proteins involved in glycolysis, ethanol fermentation and the oxidative stress response, but a decrease in proteins involved in the tricarboxylic acid cycle and the pentose phosphate shunt (Barker et al., 2012). DTT susceptibility has also been linked to plasma membrane and cell wall stability; this effect is observed most strikingly when the yeast C. albicans is treated with tetracycline, as tetracycline induces a loss of mitochondrial function and total ergosterol, which is perhaps why tetracycline is synergistic with amphotericin B (Oliver et al., 2008). High concentrations of tetracycline also render C. albicans more resistant to DTT, suggesting that alterations in the cell wall stability can be countered by decreasing the total ergosterol content of the organism (Oliver et al., 2008). Although limited phenotypes were observed when SMO mutants were exposed to cell wall-perturbing or antifungal agents, high temperature or H₂O₂, significant alterations were observed in SMO mutants when exposed to DTT, calcofluor white and low-iron conditions, suggesting an important role for these proteins in maintenance of the stress adaptation machinery (Fig. 5a–c and data not shown). These phenotypic data beg several questions regarding the specific role of Erg25 enzymes per se in mediating these phenotypes as well as the accumulated sterol intermediates. It is likely that both the enzymes themselves and the sterol intermediates contribute to the physiological impact of the fungal cell and its ability to respond to stress.

For example, a yeast two-hybrid screen has identified >140 genetic or physical interactors with Erg25 in S. cerevisiae (Sukhanova et al., 2013), including components of ergosterol biosynthesis, secretion, vesicle-mediated transport, localization, membrane lipid biosynthesis, sphingolipid biosynthesis and ceramide biosynthesis (Sukhanova et al., 2013). This evidence from S. cerevisiae and H. sapiens clearly indicates a protein and/or genetic function for Erg25/erg25 outside of ergosterol biosynthesis. It is plausible, therefore, that some of the phenotypes observed in Δerg25A and Δerg25B are due to uncharacterized interactions with non-ergosterol proteins, and not solely due to dysregulation of membrane biosynthesis.

With regard to sterol intermediates, in this study we observed elevated levels of parkeol in Δerg25B (Fig. 3a). Parkeol differs from lanosterol, 4,4,14-trimethyl-cholesta-8,24-dien-3β-ol, only in the position of a single double bond. Although the differences between parkeol and lanosterol are slight, biochemical evidence in other fungal organisms suggests that demethylation of C14, and a subsequent accumulation of C4-methyl sterols, cannot occur with the intermediate parkeol as a substrate (Venkatramesh & Nes, 1995). This finding is additionally significant as the accumulation of C14-methylated sterols is toxic when incorporated into the fungal cell membrane, as is seen with the administration of triazole antifungal drugs (Kelly et al., 1995). The mechanism used to produce ergosterol through parkeol, and the impact this altered mechanism has on fungal growth and virulence, have yet to be verified biochemically. Interestingly, erg7 mutants display altered lanosterol-to-parkeol ratios when compared with a WT erg7C (Kimura et al., 2010). These results suggest that precursor oxidosqualene cyclization can produce parkeol in a manner similar to that described for lanosterol. How alterations in erg25 result in alteration of erg7 efficiency or action are unknown at this time. Moreover, an important future direction is to determine the biological importance of alterations in ergosterol biosynthesis from episterol to ergosterol (Fig. 3c). This portion of the pathway has been described as a ‘fishing net’; when one disturbance pulls in an area, another section of the net takes up the load. More studies will be required to elucidate the requirement for individual enzymes or transcription factors in substrate acquisition and pathway flux.

Examples of the critical role of sterol intermediates themselves in eukaryotic biology are found in humans. In H. sapiens, Erg25 (SC4MOL/MSMO1) plays a critical role in the maturation of oocytes and activation of meiosis; importantly, C4-methyl sterol intermediates also regulate inflammation and cholesterol metabolism as liver X receptor ligands (Castrillo et al., 2003; Janowski et al., 1996). Mutations in SC4MOL/MSMO1 alter cytokine production, including TNF-α and IL-6 (He et al., 2011; Sukhanova et al., 2013), and silencing of SC4MOL has been linked to sensitization of refractory tumour cells to epidermal growth factor receptor inhibitors (Sukhanova et al., 2013). Whether receptors for fungal C4-methyl sterols exist and how they impact fungal physiology and/or interactions with the host is an important future research direction.

Along the lines of Erg25 and sterol intermediate biological function in A. fumigatus, a significant increase in erg25A mRNA levels in ΔsrbA was unable to fully restore hypoxic growth of this important virulence factor. To confirm that C4-methyl sterol accumulation is not solely responsible for ΔsrbA’s hypoxic growth defect would require quantification of the C4-methyl sterols in the reconstituted strain. However, the large increase in erg25A mRNA levels and the relatively modest increase in hypoxic growth of ΔsrbA is suggestive that C4-methyl sterol accumulation and/or alternation of sterol biosynthesis in general in ΔsrbA is involved in the hypoxic growth defect. Given the complexities of SrbA function, it is unclear whether full reconstitution of ΔsrbA hypoxic growth is possible through genetic means.

In conclusion, our study characterizes the individual roles of SMOs in A. fumigatus and further emphasizes the link between fungal SREBPs, ergosterol biosynthesis and stress adaptation. Although genetic deletion of erg25B resulted in no significant phenotypes under the conditions tested, save an enhanced susceptibility to the ER stress-inducing agent DTT, deletion of erg25A impacted the stress response of A. fumigatus to hypoxia and ER perturbation. In addition, SMO mutants displayed alterations in 9(11)-sterol intermediates, and a possible substrate-specific mechanism for traversing between episterol and ergosterol. In humans, C4-sterol demethylation represents a rate-limiting step in active cell proliferation, as SMOs are amongst the slowest enzymes in cholesterol biosynthesis (Bouvier et al., 2005; Rahier, 2011). It has been hypothesized that SMO activity is part of the ‘kinetic control’ of sterol biosynthesis and a mechanism to regulate the flow of carbon in the cell (Bouvier et al., 2005; Rahier, 2011). Future studies in pathogenic fungi are needed to develop a fuller understanding of the mechanisms and biological impact of sterol biosynthesis enzymes and their resulting intermediates on virulence and responses to antifungal drugs.

Abbreviations:

ER

endoplasmic reticulum

q

quantitative

RT

real-time

SMO

C4-sterol methyl oxidase

SREBP

sterol regulatory element binding protein