A fungal ABC transporter FgAtm1 regulates iron homeostasis via the transcription factor cascade FgAreA-HapX

Summary

Iron homeostasis is important for growth, reproduction and other metabolic processes in all eukaryotes. However, the functions of ATP-binding cassette (ABC) transporters in iron homeostasis are largely unknown. Here, we found that one ABC transporter (named FgAtm1) is involved in regulating iron homeostasis, by screening sensitivity to iron stress for 60 ABC transporter mutants of Fusarium graminearum, a devastating fungal pathogen of small grain cereal crops worldwide. The lack of FgAtm1 reduces the activity of cytosolic Fe-S proteins nitrite reductase and xanthine dehydrogenase, which causes high expression of FgHapX via activating transcription factor FgAreA. FgHapX represses transcription of genes for iron-consuming proteins directly but activates genes for iron acquisition proteins by suppressing another iron regulator FgSreA. In addition, the transcriptional activity of FgHapX is regulated by the monothiol glutaredoxin FgGrx4. Furthermore, the phosphorylation of FgHapX, mediated by the Ser/Thr kinase FgYak1, is required for its functions in iron homeostasis. Taken together, this study uncovers a novel regulatory mechanism of iron homeostasis mediated by an ABC transporter in an important pathogenic fungus.

Keywords:

Saccharomyces cerevisiae – Transcription factors – Homeostasis – mitochondria – DNA transcription – Fluorescence microscopy – Phosphorylation – Transcriptional control

Introduction

Iron is an essential element for growth, and can be present in various forms such as iron ions, heme and iron sulfur clusters that play critical roles in respiration, DNA synthesis and repair, ribosome biogenesis, metabolism and other cellular processes in all organisms [1–3]. In mammals, iron deficiency anemia is the most extended and common nutritional disorder in the world [4, 5]. In pathogenic fungi, the defects in iron uptake lead to decreased virulence [6, 7]. However, excess iron has the ability to generate toxic reactive oxygen species (ROS) through Fenton's reaction resulting in damage to cellular components [8]. Iron overload in liver and other organs from hepcidin regulation disorder is associated with hereditary hemochromatosis [9, 10]. Consequently, all organisms have developed tightly homeostatic regulatory mechanisms to balance uptake, consumption and storage of iron.

ATP-binding cassette (ABC) transporters that contain transmembrane domains (TMDs) and structurally conserved nucleotide-binding domains (NBDs) actively transport a wide variety of compounds across biological membranes [11]. ABC transporters play important roles in transporting compounds and regulating various physiological processes, including fatty acid metabolism, ribosome biogenesis, and mRNA translation [12, 13]. Recently, ABC transporters have been implicated in endocytosis and hyphal formation in Candida albicans [14], autophagy in human [15, 16], lifespan regulation in Drosophila [17], and the establishment of terrestrial lifestyle in plants [18]. However, our understanding of ABC transporters involved in iron homeostasis is limited. In S. cerevisiae, two ABC transporters Atm1 and Mdl1 have been found to be associated with iron homeostasis. Atm1 regulates the assembly of cytoplasmic and nucleic iron-sulfur (Fe-S) proteins might via transporting glutathione (GSH)-linked [2Fe-2S] clusters ((GS)₄-[2Fe-2S]) from mitochondria to cytosol [19–20]. Mdl1 exports the proteolytic products generated by the m-AAA protease, and the over-expression of Mdl1 partially restores the defects in Atm1 mutant [21, 22]. In addition, an ABC transporter in Mycosphaerella graminicola (MgAtr7) harboring a dityrosine/pyoverdine biosynthetic domain is required for siderophore production and subsequently modulates iron homeostasis [23].

To date, Atm1 homologs have been found to control assembly of cytoplasmic and nuclear Fe-S proteins in S. cerevisiae, Arabidopsis thaliana and Homo sapiens [24–26]. But the regulatory mechanism of iron homeostasis modulated by Atm1 is only characterized in S. cerevisiae [27–30]. In the budding yeast, the Fe-S proteins monothiol glutaredoxins Grx3/4 sense GSH-linked [2Fe-2S] clusters exported by Atm1 from mitochondria to cytoplasm [31]. Depletion of ATM1 impairs the loading of GSH-linked [2Fe-2S] clusters onto monothiol glutaredoxins Grx3/4, thus hindering the formation of the complex containing Grx3/4 and the cytosolic proteins Fra1/2. This subsequently enhances retention of the transcription factor Aft1/2 at the promoter of iron acquisition genes, therefore leading to constitutive gene activation [27–31]. Except for the budding yeast, the functions and regulatory mechanisms of Atm1 orthologs in iron homeostasis have not been documented in other organisms.

F. graminearum is an economically important plant pathogen that causes cereal scab disease worldwide [32]. In addition to yield reduction, mycotoxins such as deoxynivalenol (DON) and zearalenone (ZEA) produced by the causal agent constitute a serious threat to food security and human health [33]. F. graminearum contains many more ABC transporters than most other representative fungi from major evolutionary lineages within the fungal kingdom [34, 35]. After the screening of 60 ABC knockout mutants for sensitivity to iron stress, we found that only the FgAtm1 (Atm1 ortholog) mutant was highly sensitive, whereas the mutants of Mdl1 (FGSG_01885) and MgAtr7 (FGSG_03735) orthologs were not involved in iron regulation in F. graminearum. We therefore focused on exploring the functions of FgAtm1 in regulating iron homeostasis in F. graminearum.

In this study, we revealed that the deletion of FgATM1 impedes the activity of cytosolic Fe-S proteins nitrite reductase and xanthine dehydrogenase, which in turn induces transcription factor FgAreA, and subsequently activates the transcription factor FgHapX. The phosphorylation of FgHapX is mediated by the Ser/Thr kinase FgYak1 and is further required for the transcriptional regulation of iron-related genes. It is worth to note that this interaction between FgHapX and the monothiol glutaredoxin FgGrx4 is also required for the transcriptional activity of FgHapX, which is dramatically different from what is known in the budding yeast. Overall, results from this study reveal a regulatory mechanism of iron homeostasis mediated by FgAtm1 and the transcription factor cascade FgAreA-HapX in F. graminearum, which will help us improve the understanding of iron-homeostatic regulation in eukaryotes.

Results

Deletion of FgATM1 led to the accumulation of iron in F. graminearum

F. graminearum contains 62 putative ABC transporters. In order to explore functions of ABC transporters, we deleted each of them using a homology recombination strategy. Among 62 ABC transporter genes, 60 were deleted successfully, and two genes FGSG_07101 and FGSG_04181 are essential for F. graminearum growth [35]. To explore functions of ABC transporters in iron homeostasis, we screened these 60 deletion mutants for the sensitivity to iron stress and found that the mutant of FGSG_10911 was supersensitive to iron stress (S1 Fig), indicating that this ABC transporter may play important roles in iron homeostasis regulation. The BLAST analysis showed that FGSG_10911 is homologous to S. cerevisiae Atm1 (Fig 1A), and thus we named the gene FgAtm1. We complemented ΔFgAtm1 with FgAtm1-GFP and N-terminal mitochondrion-targeting sequence of FgAtm1 (FgAtm1^N1-111)-GFP, respectively. Subcellular localization observation revealed that FgAtm1-GFP co-localized with the mitochondrial dye MitoTracker, and the N-terminal mitochondrion-targeting sequence (http://www.cbs.dtu.dk/services/TargetP/) is thought to be responsible for its mitochondrial localization (Fig 1B and S2 Fig).

Phenotypic characterization showed that ΔFgAtm1 displayed hypersensitivity to 0.5 mM Fe²⁺ and 2 mM Fe²⁺ supplemented into minimal medium (MM) and potato dextrose agar (PDA), respectively (Fig 1C). In contrast, after treatment with iron-specific chelating agent bathophenanthroline disulfonate (BPS) at 0.3 mM, ΔFgAtm1 grew better than untreated cultures on MM and PDA (Fig 1C). Determination of intracellular iron by using fluorescent iron-binding dye FeRhoNox-1 and a colorimetric ferrozine-based assay revealed a high level of iron accumulation in both mitochondria and the whole cell of ΔFgAtm1 (Fig 1D). The iron content in mitochondria was clearly higher than that in whole cell (Fig 1D). In addition, the effect of iron stress on conidial germination was also determined. As shown in Fig 1E, 98% of the wild type conidia germinated after incubation at 28°C for 24 h in the trichothecene biosynthesis induction (TBI) medium that contains a trace amount of Fe²⁺ [36], while, ΔFgAtm1 conidia did not germinate even after 48 h under the same conditions. When 0.3 mM BPS was added into TBI to chelate Fe²⁺, 44% and 72% of ΔFgAtm1 conidia germinated after incubation for 24 and 48 h, respectively. Conidial germination in the wild type was not affected by BPS treatment (Fig 1E). To further confirm that the supersensitivity of ΔFgAtm1 to iron stress is due to the deletion of the FgATM1 gene, the mutant was complemented with a full-length wild-type FgATM1 gene amplified with the primers listed in S1 Table. The complemented strain ΔFgAtm1-C contained a single copy of FgATM1, which was inserted into the genome of ΔFgAtm1 (S3 Fig). The defects of mycelial growth, conidial germination and accumulation of iron in ΔFgAtm1 were restored to the wild-type phenotypes in ΔFgAtm1-C (Fig 1C–1E). These results strongly indicate that the lack of FgAtm1 leads to accumulation of intracellular iron in F. graminearum.

FgAtm1 regulates activity of cytosolic Fe-S proteins

To determine whether FgAtm1 regulates the assembly of cytosolic Fe-S proteins, we studied the activity of Fe-S proteins isopropyl malate isomerase (FgLeu1, FGSG_ 09589), aconitase (FgAco1, FGSG_07953), fumarase (FgFum1, FGSG_08712). S. cerevisiae Leu1 is a cytosolic Fe-S protein and catalyzes the second step in leucine biosynthesis [26]. Cytosolic and mitochondrial Fe-S proteins Aco1 and Fum1 both participate in glyoxylate shunt in cytosol and TCA cycle in mitochondria in the budding yeast, respectively [37, 38]. Enzyme activity assays showed that the activities of FgLeu1, FgAco1 and FgFum1 in the cytosol of ΔFgAtm1 were attenuated by 37, 44 and 28% respectively, when compared to those in the wild type. In contrast, the activities of FgAco1 and FgFum1 in ΔFgAtm1 mitochondria were not significantly changed (Fig 2A). Further, feeding ΔFgAtm1 with leucine (the final catalytic product of FgLeu1) or the final catalytic product of other cytosolic Fe-S proteins nitrite reductase, glutamate dehydrogenase or xanthine dehydrogenase [39, 40] also accelerated the growth of ΔFgAtm1 on MM (Fig 2B). In S. cerevisiae, GSH-linked [2Fe-2S] clusters were reported to be the substrate of Atm1, and deletion of ATM1 caused increased GSH content in the whole cell [19, 20, 41]. We therefore tested the content of GSH, and found that it was increased by 59% and 50% in mitochondria and the whole cell of ΔFgAtm1, respectively (Fig 2C). Similar to these reported in the budding yeast [19–20], the results of this study indicate that FgAtm1 also modulates the assembly of cytosolic Fe-S proteins likely via transporting GSH-linked [2Fe-2S] clusters from mitochondria into F. graminearum cytoplasm.

The deletion of FgAtm1 activates the transcription factor cascade FgAreA-HapX

To further explore the regulatory mechanism of FgAtm1 in iron homeostasis, we determined iron stress sensitivity for nine cytoplasmic Fe-S protein mutants constructed in our laboratory and found that nitrite reductase (FgNiiA, FGSG_08402) and xanthine dehydrogenase (FgXdh, FGSG_01561) mutants showed increased sensitivity to Fe²⁺ (Fig 3A and S5 Fig). Similar to what were reported in the yeasts [37, 38, 42, 43], the remaining proteins that we tested were not involved in iron stress responses. Previous studies have shown that nitrite reductase and xanthine dehydrogenase are key enzymes for non-preferred nitrogen source utilization, and are regulated by nitrogen metabolism regulator AreA in Aspergillus nidulans, F. oxysporum and F. graminearum [40, 44–47]. Nitrite reductase is responsible for nitrate utilization [40] and xanthine dehydrogenase is required for oxidizing hypoxanthine to xanthine [48]. Quantitative reverse transcription PCR (qRT-PCR) assays showed that transcription of FgAREA was induced by the deletion of FgATM1, FgNIIA (encoding nitrite reductase) or FgXDH (encoding xanthine dehydrogenase), as well as by the non-preferred nitrogen sources, NaNO₃ or hypoxanthine (Fig 3B). Surprisingly, we found that ΔFgAreA also exhibited elevated sensitivity to Fe²⁺ (Fig 3A). Therefore, we hypothesized that the deletion of FgAtm1 leads to reduced activities of FgNiiA and FgXdh, which induces overexpression of FgAreA.

To explore the role of FgAreA in regulating iron homeostasis, we first performed serial analysis of gene expression (SAGE) assay for the mutant ΔFgAtm1, and found that 56 iron-related genes were differentially expressed (>2-fold) in ΔFgAtm1 (S2 Table). Further, qRT-PCR assay confirmed that the transcription level of the transcription factor FgHAPX was dramatically increased in ΔFgAtm1 as compared to that of the wild type (S6 Fig). To understand the mechanism by which FgAreA regulates FgHAPX expression, we studied the binding ability of FgAreA to the promoter of FgHAPX in the wild type bearing FgAreA-GFP (PH-1::FgAreA-GFP) and in ΔFgAtm1 bearing FgAreA-GFP (ΔFgAtm1::FgAreA-GFP) using chromatin immunoprecipitation and quantitative PCR (ChIP-qPCR) assay. A strain transformed with GFP alone was used as a negative control. ChIP-qPCR analyses showed that enrichment of FgAreA at the FgHAPX promoter was induced by the deletion of FgATM1 as well as the treatment by NaNO₃ or hypoxanthine (Fig 3C). GFP enrichment at the FgHAPX promoter was undetectable in the negative control strain (Fig 3C). Additionally, qRT-PCR assays revealed that FgHAPX transcription was also induced with NaNO₃ or hypoxanthine treatment. Moreover the induced expression of FgHAPX upon non-preferred nitrogen source treatment was dependent on FgAreA (Fig 3D). These results indicated that FgAreA binds to the promoter of FgHAPX and regulates its transcription.

FgHapX suppresses transcription of iron-consuming genes directly and also activates transcription of iron acquisition genes by repressing another transcription factor FgSreA

To explore the function of FgHapX in iron homeostasis, we first constructed a FgHAPX deletion mutant ΔFgHapX, and tested the sensitivity of ΔFgHapX to iron stress. As shown in Fig 4A and 4B, ΔFgHapX became more sensitive to iron stress in comparison with the wild type, although ΔFgHapX did not show an obvious change in total iron content. Furthermore, we determined the content of extra- and intracellular siderophores secreted by ΔFgHapX with a chrome azurol S (CAS) assay, and found that the lack of FgHAPX caused reduced extracellular siderophore but not intracellular siderophore (Fig 4C). Similarly, qRT-PCR assays revealed that iron acquisition genes were down-regulated and iron-consuming genes were up-regulated in ΔFgHapX (Fig 4D). We knocked out FgHAPX in ΔFgAtm1, and checked whether the defects of ΔFgAtm1 were partially recovered by deletion of FgHAPX. As we expected, the double mutant ΔFgAtm1-HapX grew better and displayed decreased sensitivity to iron stress than ΔFgAtm1 (Fig 4A). Determination of iron and siderophore revealed that lack of FgHAPX in ΔFgAtm1 led to a reduced iron concentration, and decreased extra- and intracellular siderophores in comparison with those in ΔFgAtm1 (Fig 4B and 4C). Expression levels of iron acquisition genes in ΔFgAtm1-HapX were reduced and the transcription of iron-consuming genes were elevated compared to those in ΔFgAtm1 (Fig 4D).

HapX homologs in A. nidulans and A. fumigatus have been found to repress the transcription of iron-consuming genes by binding to CCAAT motif [49, 50]. The multiple EM for motif elicitation (MEME) analyses showed that the genes FgCYCA, FgHEMA, FgLYSF and FgACOA involved in the iron-consuming have the CCAAT motif (Fig 5A). Electrophoretic mobility shift assay (EMSA) further confirmed that FgHapX bound the promoters of iron-consuming genes (Fig 5B). Iron acquisition genes contained the GATA, but not CCAAT, motif in their promoters (Fig 5A), indicating that other regulator(s) modulates the transcription of iron uptake genes directly in F. graminearum. In A. fumigatus, HapX activates the expression of siderophore-mediated iron uptake genes via transcriptional repression of SreA that suppresses the transcription of iron acquisition genes via binding to the GATA motif in their promoter [50, 51]. The MEME analysis and EMSA assay showed that FgHapX could bind the FgSREA (SreA homolog) promoter (Fig 5A and 5B). Moreover, the qRT-PCR assay revealed that deletion of FgHAPX led to elevated transcription of FgSREA (Fig 5C). We further obtained a FgSREA deletion mutant ΔFgSreA, and found that ΔFgSreA displayed increased sensitivity to iron stress (Fig 5D). The deletion of FgSREA caused iron accumulation, and increased extra- and intracellular siderophores (Fig 5E and 5F), and qRT-PCR assays showed that deletion of FgSREA caused elevated expression of iron acquisition genes (Fig 5G). These results indicated that FgHapX represses the transcription of FgSreA, and subsequently activates transcription of iron acquisition genes.

FgHapX interacts with FgGrx4 to regulate iron homeostasis

To explore the regulatory mechanism of FgHapX in iron homeostasis, we performed a yeast two-hybrid (Y2H) screen of F. graminearum cDNA library, and found 50 potential FgHapX-interacting proteins (S3 Table), including the monothiol glutaredoxin FgGrx4 that is homologous to S. cerevisiae Grx3/4. Furthermore, Y2H, Co-IP and BiFC assays revealed that FgGrx4 interacted with FgHapX in the nucleus and the interaction was independent of FgAtm1 (Fig 6A–6D). Moreover, Y2H and BiFC assays showed that FgGrx4 interacted with FgHapX through its GRX domain but not its TRX domain (Fig 6B and 6D). In S. cerevisiae, lack of Grx3/4 leads to constitutive expression of iron acquisition genes, which contributes to iron accumulation in cells [52, 53]. We therefore generated a FgGRX4 deletion mutant, ΔFgGrx4, and found that ΔFgGrx4 displayed increased sensitivity to iron stress although it did not exhibit an obvious alteration in the total iron content (Fig 6E and S7A Fig). To explore the effect of FgGrx4 on FgHapX functions, we determined the quantity and localization of FgHapX in ΔFgGrx4, and found that FgGRX4 deletion did not cause an obvious change in the localization and quantity of FgHapX (S7B and S7C Fig). However, similar to the FgHapX deletion, the deletion of FgGRX4 led to significantly decreased expression of iron acquisition genes, and increased transcription of iron-consuming genes (Fig 6G). These results indicated that FgGrx4 is required for the transcriptional activity of FgHapX in F. graminearum.

FgHapX is phosphorylated by the Ser/Thr kinase FgYak1

In eukaryotes, phosphorylation of transcription factors frequently has been found to regulate their activities. Phosphoproteome assay showed that FgHapX contains two predicted Ser residues at 245 and 338 sites that may be subject to phosphorylation (S8 Fig). To confirm the function of these two residues, we constructed a strain containing a constitutive dephosphorylated FgHapX isoform. Briefly, the two phosphorylated Ser residues were replaced by alanine, the mutated FgHapX^S245A/S338A was transformed into ΔFgHapX and the resulting strain was designated as ΔFgHapX-C^S245A/S338A. Next, we performed a phos-tag assay to detect the phosphorylation level of FgHapX in ΔFgHapX-C and in ΔFgHapX-C^S245A/S338A. As shown in Fig 7A, the dephosphorylated level of FgHapX in ΔFgHapX-C^S245A/S338A was significantly higher than that in ΔFgHapX-C. To further explore the function of FgHapX phosphorylation, we determined the sensitivity of ΔFgHapX-C^S245A/S338A to iron stress. As shown in Fig 7B and 7C, similar to ΔFgHapX, ΔFgHapX-C^S245A/S338A still remained highly sensitivity to iron stress. Consistently, the qRT-PCR assays showed the expression levels of iron acquisition genes were reduced and the transcription of iron-consuming genes was elevated in ΔFgHapX-C^S245A/S338A (Fig 7D). In addition, these mutations did not change the quantity and localization of FgHapX (S9A and S9B Fig). Collectively, these results indicate that phosphorylation of FgHapX is required for regulating expression of iron-related genes.

To identify the potential kinase that phosphorylates FgHapX, we screened 96 kinase mutants and found that the mutant of Ser/Thr protein kinase FgYak1 (FGSG_05418) showed dramatically increased sensitivity to iron stress (Fig 7B and S10 Fig). Furthermore, co-immunoprecipitation (Co-IP) confirmed that FgYak1 interacted with FgHapX (Fig 7E). Immunofluorescence assay also revealed that FgYak1 interacted with FgHapX in the nucleus (Fig 7F). The qRT-PCR assays showed that, similar to those in ΔFgHapX and ΔFgHapX-C^S245A/S338A, the transcription levels of iron acquisition genes were reduced and those of iron-consuming genes were elevated in ΔFgYak1 (Fig 7D). Importantly, the phos-tag assay revealed that the dephosphorylated level of FgHapX in ΔFgYak1 was higher than that in ΔFgHapX-C (Fig 7A). Meanwhile, the phos-tag assays showed that the phosphorylated levels of FgHapX in ΔFGSG_13318, ΔFGSG_00408, ΔFGSG_10381, ΔFGSG_06832, ΔFGSG_05734 or ΔFGSG_11812 were similar with that in ΔFgHapX-C, although these kinase mutants also showed elevated sensitivity to iron stress (S10A–S10D Fig). In addition, deletion of FgYak1 did not alter the quantity and localization of FgHapX (S9A and S9B Fig). Collectively, these results indicated that FgYak1 phosphorylates FgHapX in F. graminearum.

Discussion

In S. cerevisiae, the Atm1-mediated iron regulation has been well characterized. The depletion of Atm1 impedes the loading of GSH-linked [2Fe-2S] clusters onto monothiol glutaredoxins, subsequently disrupting formation of the Grx3/4-Fra1/2 complex, which results in the failure of Aft1/2 dissociation from the promoters of iron acquisition genes [27–30]. The iron acquisition genes are therefore activated constitutively in the Atm1-depleted cells. In this study, we found that lack of FgAtm1 also leads to an overload of intracellular iron. However, the regulation mechanism of iron homeostasis mediated by FgAtm1 in F. graminearum is different from what is known in S. cerevisiae. We uncovered that deletion of FgAtm1 impedes the activity of cytosolic Fe-S proteins nitrite reductase and xanthine dehydrogenase, which conversely activates the nitrogen metabolism regulator FgAreA (Fig 8). Subsequently, FgAreA activates the transcription of repressor FgHapX via binding the FgHAPX promoter (Fig 8). Moreover, we found that FgHapX directly represses the transcription of iron-consuming genes, but also activates the expression of iron acquisition genes indirectly via suppressing the transcription of another repressor FgSREA (Fig 8). It is worthy to note that S. cerevisiae does not contain a HapX ortholog, and F. graminearum and other filamentous fungi do not have the yeast Aft1/2 orthologs. These results indicate that the regulatory networks of iron homeostasis can be distinct in different fungi.

AreA belongs to the GATA nitrogen regulator and is required for the transcription of genes responsible for the utilization of non-preferred nitrogen sources in several fungi [54]. Previous studies have found that the transcription of AREA is induced by nitrogen starvation or the treatment with nitrate in A. nidulans [55], F. graminearum [56, 57] and Fusarium fujikuroi [58]. In this study, the lack of FgAtm1 compromised the activity of nitrite reductase FgNiiA and xanthine dehydrogenase FgXdh resulting in utilization defects of the non-preferred nitrogen sources nitrite and hypoxanthine, and subsequently induced the transcription of FgAREA. This finding indicates that iron metabolism is able to affect nitrogen utilization via the Fe-S proteins in the filamentous fungus F. graminearum.

HapX is an important transcriptional repressor of intracellular iron homeostasis in filamentous fungi [59]. In A. fumigatus, HapX not only represses genes involved in iron-consuming pathways to spare iron, but also activates iron acquisition genes to acquire iron via suppressing another repressor SreA during iron starvation [50], which is consistent with our finding. However deletion of HapX homologs does not change the transcription of iron acquisition genes in A. nidulans, F. oxysporum and C. albicans [49, 60, 61]. Previous studies have found that the transcription factor SreA may also represses the expression of HapX during iron overload [49, 62]. In Cryptococcus neoformans, carbon metabolism regulator Mig1 promotes the transcription of HAPX under low-iron conditions [63]. In this study, we found however that FgHapX is regulated by the GATA transcription factor FgAreA. To our knowledge, this is the first observation that the iron regulator HapX is regulated by a nitrogen metabolism regulator AreA in filamentous fungi.

In the current study, we also found that the functions of Grx4 orthologs vary dramatically in F. graminearum and yeasts. First, monothiol glutaredoxin FgGrx4 is required for the transcriptional activity of FgHapX via its interaction with FgHapX in F. graminearum. In yeasts, however, Grx4 homologs combine with the GSH-linked [2Fe-2S] clusters and then interact with transcription factors S. cerevisiae Aft1/Aft2 or Schizosaccharomyces pombe Php4/ Fep1 to disassociate these factors from the promoters of iron regulation genes [28–30, 64–66]. Second, the interaction of FgGrx4 and FgHapX is independent of the presence of FgAtm1, and the TRX domain of FgGrx4 doesn’t interact with FgHapX. In S. cerevisiae, the interaction of Grx4 and Aft1/2 is dependent on Atm1 [30]. In S. pombe, the interaction of GRX domain of Grx4 with Php4 or Fep1 is dependent on iron conditions, whereas the TRX domain continuously binds to Php4 or Fep1 [65, 66]. Third, similar to FgHAPX deletion, deletion of FgGRX4 did not change intracellular iron content (S7A Fig). However, lack of Grx3/4 leads to iron accumulation in S. cerevisiae cells [52]. Fourth, deletion of FgGRX4 caused the down-regulation of iron acquisition genes in F. graminearum. However, deletion of GRX3 or GRX4 causes constitutive expression of iron acquisition genes in S. cerevisiae [52, 53]. In S. pombe, GRX4 disruption causes constitutive transcription of iron acquisition genes regulated by Fep1, or constitutive transcription of iron-consuming genes regulated by Php4 [64, 66]. Taken together, these results indicate that the function of FgGrx4 in F. graminearum is dramatically different from that of Grx4 in the yeasts.

Phosphorylation of transcription factors mediated by various kinases has been found to modulate their localization, protein accumulation and DNA binding ability in eukaryotic organisms [67]. In mammals, phosphorylation of the organismal lifespan-related transcription factor FOXO by serum and glucocorticoid-induced kinase (SGK) results in the exclusion from the nucleus and repression of transcriptional activity [68]. In A. thaliana, multisite light-induced phosphorylation of phy-interacting basic Helix Loop Helix (bHLH) transcription factor PIF3 causes its degradation [69]. In the fission yeast, phosphorylation of sterol biosynthesis regulator SpSre1 mediated by casein kinase Hhp2 reduces its protein quantity by accelerating its degradation [70]. Whereas in Lotus japonicus, phosphorylation of a root nodule development-associated transcriptional activator CYCLOPS by calcium- and calmodulin-dependent kinase (CCaMK) increases its DNA binding activity at the target gene promoters [71]. In this study, we discovered that phosphorylation of FgHapX is required for its transcription activity, but not for its quantity and localization (Fig 6D, S9A and S9B Fig). Furthermore, we identified that FgHapX is subject to phosphorylation mediated by the kinase FgYak1. Previous studies have reported that the Ser/Thr protein kinase Yak1 controls cell growth in response to glucose depletion by negatively regulating the cAMP-PKA pathway in S. cerevisiae, and regulates the emergence and maintenance of hyphal growth of C. albicans [72, 73]. To our knowledge, it is the first report on involvement of a kinase (Yak1) in regulating iron homeostasis by phosphorylation of a transcription factor in fungi.

In A. thaliana, Atm1 ortholog Atm3 was found to regulate the assembly of cytoplasmic molybdenum cofactor (Moco) proteins, besides Fe-S proteins [Bernard et al., 2009]. The precursor of Moco cyclic pyranopterin monophosphate (cPMP) that is synthesized in mitochondria was reported as another substrate of Atm3 [25, 74]. In the current study, we also found that F. graminearum FgAtm1 modulates the assembly of Moco protein nitrate reductase (FgNiaD) (S11 Fig). Further, we also found that F. graminearum FgAtm1 modulates mitochondrial function and redox balance besides iron homeostasis, which are in agreement with the studies in these reports of S. cerevisiae and C. neoformans [41, 75, 76]. The ΔFgAtm1 mutant displayed decreased sensitivity to the mitochondrial respiratory inhibitors diphenylene iodonium (DPI) and rotenone (complex I) and antimycin A (complex III) (S12A Fig). The mutant also exhibited increased reactive oxygen species (ROS) content and elevated sensitivity to hydrogen peroxide (H₂O₂) (S12B–S12D Fig), since mitochondrial respiratory complexes I and III are known to be major generators of ROS in eukaryotic cells [77]. In addition, phenotypic determination showed that deletion of FgATM1 led to the defects in asexual and sexual development, virulence and secondary metabolite production (S13 Fig). The phenotypic defects might result from the imbalance of nutrient, iron and redox and impaired mitochondrial functions in ΔFgAtm1.

Materials and methods

Fungal strains and sensitivity determination

F. graminearum wild-type strain PH-1 was used as a parental strain for transformation experiments in this study. Mycelial growth of the wild type and the resulting transformants were assayed on potato dextrose agar (PDA) or minimal medium (MM) as described previously [78, 79]. To determine sensitivity to iron stress, 5-mm mycelial plugs of each strain taken from a 3-day-old colony edge were inoculated on PDA or MM supplemented without/with Fe²⁺, H₂O_2, or catalytic product of each Fe-S protein, and then incubated at 25°C for 3 days in the dark. Three biological replicates were used for each strain and each experiment was repeated three times independently.

Construction of gene replacement, GFP-, Flag-, CYFP-, NYFP- and mCherry- fusion cassettes

The double-joint PCR approach [80] was used to generate the gene replacement construct for each target gene. Briefly, for each gene, 5’ and 3’ flanking regions were amplified with the primer pairs listed in S1 Table and the resulting amplified sequences were then fused with the hygromycin resistance gene cassette (HPH) driven by the constitutive trpC promoter which was amplified from the pBS-HPH1 vector [81]. Protoplast transformation of F. graminearum was carried out using the protocol described previously [82]. Putative gene deletion mutants were identified by PCR assays with relevant primers (S1 Table) and the FgATM1 deletion mutation was further confirmed by Southern hybridization assays (S3 Fig).

To construct the FgAtm1-GFP cassette, FgATM1 containing the promoter region and open-reading frame (without the stop codon) was amplified with the relevant primers (S1 Table). The resulting PCR products were co-transformed with XhoI-digested pYF11 containing a geneticin resistance gene (NEO) [83] into the yeast strain XK1-25 [84] using the Alkali-Cation Yeast Transformation Kit (MP Biomedicals, Solon, USA) to generate the recombined FgAtm1-GFP fusion vector. Subsequently, the FgAtm1-GFP fusion vector was recovered from the yeast transformant using the Yeast Plasmid Kit (Solarbio, Beijing, China) and then transferred into Escherichia coli strain DH5α for amplification. Using the similar strategy, FgYak1 (FGSG_05418)—and FgTri1 (FGSG_00071) -GFP fusion cassettes were constructed. Using the similar strategy, FgGrx4 (FGSG_01317)—and FgYak1-Flag fusion cassettes were constructed by co-transformation with XhoI-digested PHZ126 vector. Similarly, FgHapX-CYFP and FgGrx4-NYFP fusion cassettes were constructed by co-transformation with XhoI-digested PHZ68 and PHZ65 vectors, respectively.

The double-joint PCR approach [80] was also used to construct FgHapX-, FgHapX^S245A/S338A- and FgLeu1 (FGSG_09589)-mCherry cassettes. Briefly, the target gene containing the promoter region and open-reading fragment and the geneticin resistance gene (NEO) were amplified, and then fused with mCherry fragment. Before protoplast transformation, each fusion construct was verified by DNA sequencing. The transformation of F. graminearum was carried out using the previously described protocol [82]. All the mutants generated in this study were preserved in 15% glycerol at −80°C.

Determination of iron content and siderophore

For determination of mitochondrial and total iron content, About 50 mg of fresh mycelia were lysed by 2% cellulase (Ryon Biological Technology CO, Ltd, Shanghai, China), 2% lysozyme (Ryon Biological Technology CO, Ltd, Shanghai, China) and 0.2% driselase from Basidiomycetes sp. (Sigma, St. Louis, MO, USA) for 4–6 hours. After filtration with funnel and filter paper, the filtrate was centrifuged at 5000 g at 4°C for 10 min. The protoplast was used for total iron determination and mitochondrial extraction with a Cell Mitochondria Isolation Kit (Beyotime Industrial Co., Ltd., Shanghai, China). Iron content was determined using the colorimetric ferrozine-based assay previously reported, with ferrozine as chelator and ferric chloride as a standard [28, 29, 85]. Briefly, aliquots (100 μl) of cell lysates were mixed with 100 μl of 10 mM HCl (the solvent of the iron standard FeCl₃), and 100 μl of the iron-releasing reagent (a freshly mixed solution of equal volumes of 1.4 M HCl and 4.5% (w/v) KMnO₄ in H₂O). The mixtures were incubated within a fume hood at 60°C for 2 h. After the mixtures had cooled to room temperature, 30 μl of the iron-detection reagent (6.5 mM ferrozine, 6.5 mM neocuproine, and 2.5 M ammonium acetate and 1 M ascorbic acid) was added to each tube. After 30 min, 200 μl of the solution in each tube was transferred into a well of a 96-well plate and the absorbance was measured at 550 nm on a microplate reader. The linear range of the ferrozine assay is from 0.2 to 30 nmol.

To determine the content of extra- and intracellular siderophores, each strain was cultured in CM for 36 hours, then transferred to MM lack of FeSO₄ and amended 0.3 mM iron-specific chelating agent bathophenanthroline disulfonate (BPS) at 25°C for 8 hours. After filtration with funnel and filter paper, the supernatant was used for determining the content of extracellular siderophore, and the mycelia was used for determining the content of intracellular siderophore. Finely ground mycelia (50 mg) were resuspended in 50 mM potassium phosphate buffer. After vigorous vortexing, the cellular debris was pelleted, and the supernatant was determined for the content of intracellular siderophore. An aliquot of supernatant (1 ml) was mixed with 1 ml chrome azurol S (CAS) assay solution that was prepared. After incubation in the dark for 1 hour at room temperature, the absorbance of each sample was measured at 630 nm and the relative content of siderophore was calculated according to previous study [86].

Conidiation, pathogenicity and DON biosynthesis assays

For the conidiation assay, fresh mycelia (50 mg) of each strain were inoculated in a 50-ml flask containing 20 ml of carboxymethyl cellulose (CMC) liquid medium. The flasks were incubated at 25°C for 4 days in a shaker (180 rpm). Subsequently, the number of conidia in each flask was determined using a hemacytometer. Three biological replicates were used for each strain and each experiment was repeated three times independently.

Virulence assays on wheat spikelets, corn silks, and wheat seedling leaves were conducted as described previously [35]. For virulence on wheat spikelets, a 10 μl aliquot of conidial suspension (10⁵ conidia/ml) was injected into a floret in the central section spikelet of a single flowering wheat head of susceptible cultivar Jimai 22. Fifteen days after inoculation, the infected spikelets in each inoculated wheat head were recorded. For virulence on corn silks and wheat seedling leaves, a 5-mm mycelial plug of each strain was inoculated on the middle of corn silks and wheat seedling leaves, and then cultured for 4 days. Ten replicates were used for each strain and each experiment was repeated three times independently. To determine DON production, each strain was grown in liquid trichothecene biosynthesis induction (TBI) medium at 28°C for 3 d in a shaker (150 rpm) in the dark. A DON Quantification Kit (Wise Science, Zhenjiang, China) was used to quantify the DON production for each sample. The experiment was repeated three times.

Microscopic examinations

The colony edge of each strain cultured on PDA plates at 25°C for 7 days was examined with a Leica TCS SP5 imaging system (Leica Microsystems, Wetzlar, Germany). Morphology of conidia cultured in CMC liquid medium for 4 days was observed with a Leica TCS SP5 imaging system after staining with the cell wall-damaging agent calcofluor white (CFW) at 10 μg/ml. Germinating conidia in TBI liquid medium or iron-depleted TBI liquid medium were observed with a Leica TCS SP5 imaging system. Fluorescence signals were examined with a Zeiss LSM780 confocal microscope (Gottingen, Niedersachsen, Germany). For the observation of proteins tagged with GFP, mCherry or YFP, each strain was cultured in CM at 25°C for 24 h in a shaker (180 rpm) before staining with MitoTracker or 4’,6-diamidino-2-phenylindole (DAPI) [87]. For examination of iron content, each strain was stained with 5 μM FeRhoNox-1 (Goryo Chemical, Inc., Sapporo, Japan) [88–91] for 1 hours after culture in complete medium (CM) for 36 hours. For the observation of FgTri1-GFP, each strain was cultured on wheat seedling leaves at 25°C with 100% humidity for two days.

Determination of enzyme activity and ROS content

The wild type, ΔFgAtm1 and the ΔFgAtm1-C were grown in CM for 36 hours, and then were harvested after rinsing 3 times with sterile water. To determine FgLeu1 activity in cytosol, 50 mg of finely ground mycelia was resuspended in 1 ml of lysis buffer (1 M Tris-HCl pH 7.4, 1 M NaCl, 0.5 M EDTA, 1% Triton 100) for 10 min. The lysate was centrifuged at 14,000 g for 10 min at 4°C. Each protein sample (50 μg) was used for FgLeu1 activity determination. DL-threo-3-isopropylmalic acid (Wako Pure Chemical Industries, Ltd, Japan) was used as the substrate and β-isopropylmalate formation was measured by monitoring absorbance at 235 nm for 10 minutes [92, 93]. To determine FgAco1 and FgFum1 activity in cytosol and mitochondrion, the separation of mitochondrion and cytosol (without nuclei) was conducted according to the above method described for iron content determination. Each protein sample (50 μg) was used for FgAco1 activity determination. Cis-aconitic acid (Sigma-Aldrich, St. Louis, MO, USA) was used as the substrate and monitor absorbance at 340 nm for 2–4 minutes [92]. FgFum1 activity of each protein sample (50 μg) was determined with a Fumarase Specific Activity Assay Kit ab110043 (Abcam, Cambridge, UK) according to the manufacturer’s instruction. These experiments were repeated three times independently.

To examine ROS content, fresh mycelia of each strain grown in CM liquid medium for 36 hours were harvested after rinsing 3 times with sterile water. 50 mg of finely ground mycelia were resuspended in 1 ml of extraction buffer. After homogenization with a vortex shaker, the lysate was centrifuged at 14,000 g in a microcentrifuge for 10 min at 4°C. An aliquot of 10 μl supernatant was used for ROS determination with ROS ELISA Kit (Tong Wei Industrial Co., Ltd., Shanghai, China). Meantime, to examine ROS, each strain cultured on PDA plates at 25°C for 3 days was stained with 0.05% (wt/vol) nitroblue tetrazolium (NBT) for 2 h. These experiments were all repeated three times.

RNA extraction and quantitative reverse transcription PCR (qRT-PCR)

Total RNA isolation from mycelia of each sample with the TaKaRa RNAiso Reagent (TaKaRa Biotechnology, Dalian, China), and reverse transcription was performed with a HiScript II 1st Strand cDNA Synthesis Kit (Vazyme Biotech, Nanjing, China). The expression level of each gene was determined by qRT-PCR with HiScript II Q RT SuperMix (Vazyme Biotech, Nanjing, China). The expression of the FgACTIN gene was performed as a reference. The experiment was repeated three times.

Chromatin immunoprecipitation (ChIP)-qPCR analyses

ChIP was performed as previously described [94, 95] with additional modifications. Briefly, fresh mycelia were cross-linked with 1% formaldehyde for 15 min and then stopped with 125 mM glycine. The cultures were ground with liquid nitrogen and re-suspended in the lysis buffer (250 mM, HEPES pH 7.5, 150 mM NaCl, 1 mM EDTA, 1% Triton, 0.1% DeoxyCholate, 10 mM DTT) and protease inhibitor (Sangon Co., Shanghai, China). The DNA was sheared into ~300 bp fragments with twenty pulses of 10 s and 20 s of resting at 35% amplitude (Qsonica*sonicator, Q125, Branson, USA). After centrifugation, the supernatant was diluted with 10×ChIP dilution buffer (1.1% Triton X-100, 1.2 mM EDTA, 16.7 mM Tris–HCl, pH 8.0 and 167 mM NaCl). Immunoprecipitation was performed using monoclonal anti-GFP ab290 (Abcam, Cambridge, UK) antibody together with the protein A agarose (Santa Cruz, CA, USA) respectively. DNA was immunoprecipitated by ethanol after washing, eluting, reversing the cross-linking and digesting with proteinase K. Further, ChIP-enriched DNA was used for quantitative PCR analysis using SYBR green I fluorescent dye detection with the relative primers (S1 Table). Relative enrichment of each gene was determined by quantitative PCR and calculated first by normalizing the amount of a target cDNA fragment against that of FgACTIN as an internal control, and then by normalizing the value for the immunoprecipitated sample against that for the input. The ChIP-qPCR was independently repeated three times.

Electrophoretic mobility shift assay (EMSA) assays

The cDNA encoding the N terminal 1–230 amino acids of FgHapX (FgHapX^N1–230) was amplified and cloned into pGEX-4T-3 vector to generate GST-tagged protein. The resulting construct was transformed into the E. coli strain BL21 (DE3) after verifying the cDNA sequence. The recombinant GST-FgHapX^N1–230 was purified by Ni sepharose beads and eluted by reduced glutathione. Promter DNAs were amplified using the primers in Table S1. For EMSA, Reaction mixtures containing purified GST -FgHapX^N1–230, promter DNAs and 10×Binding buffer (100 mM Tris-HCl (PH 7.5), 0.5 M NaCl, 10 mM DTT, 10 mM EDTA, 50% glycerol) were incubated for 20 min at 25°C. The purified GST was used as negative controls. The reactions were electrophoresed on 1.2% agarose gel in 0.5×TAE for 45 min in 80 V under low temperature. Signals were detected in J3-3000 imaging system after dying DNA dye ethidium bromide (EB) for 15 min. The experiment was conducted independently three times.

Yeast two-hybrid (Y2H) assays

To construct plasmids for Y2H analyses, the coding sequence of each tested gene was amplified from the cDNA of PH-1 with primer pairs indicated in S1 Table. The cDNA of each gene was inserted into the yeast GAL4-binding domain vector pGBKT7 (Clontech, Mountain View, CA, USA) and GAL4 activation domain vector pGADT7 (Clontech, Mountain View, CA, USA), respectively. The pairs of Y2H plasmids were cotransformed into S. cerevisiae strain Y2H Gold lithium acetate/single-stranded DNA/polyethylene glycol transformation protocol. In addition, a pair of plasmids, pGBKT7-53 and pGADT7 and another pair of plasmids, pGBKT7-Lam and pGADT7, served as a positive control and negative controls, respectively. Transformants were grown at 30°C for 3 d on synthetic medium (SD) lacking Leu and Trp, and then transferred to SD stripped of Ade, His, Trp and Leu to assess binding activity. Three independent experiments were performed to confirm yeast two-hybrid assay results.

To search for FgHapX-interacting proteins, we performed Y2H screens. FgHapX was cloned into the yeast vector pGBKT7. A F. graminearum cDNA library was constructed in the Y2H vector pGADT7 using total RNA extracted from mycelia and conidia. The Y2HGold that was co-transformed with the cDNA library as well as FgHapX-pGBKT7 were directly selected using SD-Trp-Leu-His. Approximately 150 potential yeast transformants containing cDNA clones interacting with FgHapX were further confirmed in selection medium SD-Trp-Leu-His-Ade.

Western blotting assay

For western blotting assay, protein samples of strains were prepared and extracted as described previously [96]. Proteins separated on the SDS-PAGE gel were transferred onto a polyvinylidene fluoride membrane with a Bio-Rad electroblotting apparatus. The polyclonal anti-Flag A9044 (Sigma, St. Louis, MO, USA), monoclonal anti-GFP ab32146 (Abcam, Cambridge, MA, USA) and monoclonal anti-mCherry ab125096 (Abcam, Cambridge, UK) antibodies were used at a 1:2000 to 1:10000 dilution for immunoblot assays. The samples were also detected with the monoclonal anti-GAPDH antibody EM1101 (Hangzhou HuaAn Biotechnology co., Ltd.) as a reference. Incubation with a secondary antibody and chemiluminescent detection were performed as described previously [97]. The experiment was repeated three times independently.

Co-immunoprecipitation (Co-IP) assays

The mCherry and Flag fusion constructs were verified by DNA sequencing and transformed in pairs into PH-1. Transformants expressing the fusion constructs were confirmed by western blot analysis. In addition, the transformants bearing a single fusion construct were used as references. For Co-IP assays, total proteins were extracted and incubated with the anti-Flag (Abmart, Shanghai, China) agarose overnight at 4°C as described previously [97]. Proteins eluted from agarose were analyzed by western blotting detection with the monoclonal anti-mCherry ab125096 (Abcam, Cambridge, UK) antibody. The protein samples were also detected with monoclonal anti-GAPDH antibody EM1101 (Hangzhou HuaAn Biotechnology co., Ltd.) as a reference.

Bimolecular fluorescence complementation (BiFC) assay

The FgHapX-CYFP and FgGrx4-NYFP fusion constructs were generated by cloning the related fragments into pHZ68 and pHZ65 vectors, respectively. FgHapX-CYFP and FgGrx4-NYFP constructs were co-transformed into PH-1 and ΔFgAtm1. In addition, a pair of constructs, FgHapX-CYFP and NYFP and another pair of constructs, FgGrx4-NYFP and CYFP were used as negative controls. Transformants resistant to both hygromycin and zeocin were isolated and confirmed by PCR. YFP signals were examined with a Zeiss LSM780 confocal microscope (Gottingen, Niedersachsen, Germany).

Phosphoproteome assay

Proteins of the wild-type PH-1 were extracted in the lysis buffer (8 M urea, 50 mM Tris 8.0, 1% NP40, 1% sodium deoxycholate, 5 mM dithiothreitol, 2 mM EDTA, 30 mM nicotinamide, 3 μm trichostatin A, 1% protease inhibitor Cocktail and 1% phosphatase inhibitor cocktail). For trypsin digestion, the protein sample was diluted in 0.1 M TEAB (triethylammonium bicarbonate), and digested with 1/25 trypsin (Promega, Madison, USA) for 12 h at 37°C. The digestion was terminated with 1% TFA (trifluoroacetic acid), and the resulting peptides were cleaned with a Strata X C18 SPE column (Phenomenex, Torrance, USA) and vacuum-dried in a scanvac maxi-beta (Labogene, Lynge, Denmark). Then, the resulting peptides (2 mg per sample) were reconstituted in 120 μl 0.5 M TEAB, and treated with a TMTsixplex label reagent kit (Pierce, Idaho, USA). Both fractionations were performed with an XBridge Shield C18 RP column (Waters, Milford, USA) in a LC20AD HPLC system (Shimadzu, Kyoto, Japan). For phosphorylation enrichment, peptides were dissolved in 80% ACN/6% TFA and then incubated with IMAC-Ti4+ beads (Sachtopore, Sachtleben Chemie, Germany) at room temperature. The beads were washed once with 50% ACN/0.5%TFA/200 mM NaCl and once with 50% ACN/0.1% TFA. The bound peptides were then eluted with 10% NH₄OH and 80% ACN/2% FA (formic acid). All of the eluted fractions were combined, vacuum-dried and cleaned with C18 ZipTips (Millipore, Billerica, USA) according to the manufacturer’s instructions, followed by LC-MS/MS analysis according to a previous study [98]. The database search and bioinformatics analyses were performed as described previously [99].

Phos-tag assay analysis

For Phos-tag assay, the FgHapX-mCherry fusion construct was transferred into the ΔFgHapX and ΔFgYak1 mutants and the FgHapX^S245A/S338A-mCherry fusion construct was transferred into the ΔFgHapX mutant. Protein samples were prepared and extracted as described above. Each resulting protein sample was loaded on 8% SDS-polyacrylamide gels prepared with 25 μM Phos binding reagent acrylamide (APE×BIO, F4002) and 100 μM ZnCl₂. Gels were electrophoresed at 20 mA/gel for 3–5 h. Prior to protein transfer, gels were first equilibrated three times in transfer buffer containing 5 mM EDTA for 5 min and further equilibrated in transfer buffer without EDTA for 5 min for two times. Protein transfer from the Zn²⁺-phos-tag acrylamide gel to the PVDF membrane was performed for 4–5 h at 100 V on ice, and finally the membrane was analyzed by western blotting with the monoclonal anti-mCherry ab125096 (Abcam, Cambridge, UK) antibody.

Immunofluorescence assays

FgYak1-Flag and FgHapX-mCherry were co-transformed into the wild type. The nuclei of corresponding strains were extracted as described previously [100] and then fixed on the polylysine slides. Slides were incubated with the polyclonal anti-Flag antibody A9044 (Sigma, St. Louis, MO) and monoclonal anti-mCherry ab125096 (Abcam, Cambridge, UK) antibody at 1:500 dilution, respectively, followed by the secondary antibodies Andy Fluor 594 goat anti-mouse L119A (red fluorescence) (GeneCopoeia, Maryland, US) and Andy Fluor 488 goat anti-rabbit L110A (green fluorescence) (GeneCopoeia, Maryland, US) at 1:300 dilution. Nuclei were stained with DAPI for 15 minutes before fluorescence observation.

Supporting information

S1 Fig [a]
Sensitivity of 60 ABC transporter deletion mutants of . to FeSO.

S2 Fig [a]
Subcellular localization of FgAtm1- or FgAtm1-GFP in the complemented strain.

S3 Fig [a]
Construction and identification of gene deletion and complemented mutants.

S4 Fig [a]
The lack of FgAtm1 did not alter the localization and protein quantity of FgLeu1-mCherry.

S5 Fig [tif]
Seven potential cytoplasmic Fe-S proteins are not involved in response to iron stress in . .

S6 Fig [tif]
Relative transcription level of in the wild type and ΔFgAtm1 growth in CM for 1 day.

S7 Fig [a]
Deletion of FgGrx4 did not alter intracellular iron concentration, the localization, or protein quantity of FgHapX.

S8 Fig [a]
Mass spectrum data of two identified phosphorylated sites in FgHapX.

S9 Fig [a]
The phosphorylation of FgHapX did not alter its localization or protein quantity.

S10 Fig [a]
Sensitivity determination of 96 kinase deletion mutants of . to FeSO.

S11 Fig [tif]
FgAtm1 regulates the assembly of cytosolic Moco protein nitrate reductase (FgNiaD).

S12 Fig [a]
FgAtm1 modulates mitochondrial function and intracellular redox balance.

S13 Fig [a]
The deletion of led to defects in asexual and sexual development, virulence and secondary metabolite production.

S1 Table [docx]
The primers used in this study.

S2 Table [docx]
Serial analysis of gene expression (SAGE) for identification of differential expression (>2-folds) genes in deletion mutant as compared with those in the wild type.

S3 Table [docx]
A list of putative FgHapX-interacting proteins identified by the yeast two-hybrid assay.