The Regulatory Hub of Siderophore Biosynthesis in the Phytopathogenic Fungus Alternaria alternata

A GATA zinc finger-containing repressor (AaSreA) suppresses siderophore biosynthesis in the phytopathogenic fungus Alternaria alternata under iron-replete conditions. In this study, targeted gene deletion revealed two bZIP-containing transcription factors (AaHapX and AaAtf1) and three CCAAT-binding proteins (AaHapB, AaHapC, and AaHapE) that positively regulate gene expression in siderophore production. This is a novel phenotype regarding Atf1 and siderophore biosynthesis. Quantitative RT-PCR analyses revealed that only AaHapX and AaSreA were regulated by iron. AaSreA and AaHapX form a transcriptional feedback negative loop to regulate iron acquisition in response to the availability of environmental iron. Under iron-limited conditions, AaAtf1 enhanced the expression of AaNps6, thus playing a positive role in siderophore production. However, under nutrient-rich conditions, AaAtf1 plays a negative role in resistance to sugar-induced osmotic stress, and AaHapX plays a negative role in resistance to salt-induced osmotic stress. Virulence assays performed on detached citrus leaves revealed that AaHapX and AaAtf1 play no role in fungal pathogenicity. However, fungal strains carrying the AaHapB, AaHapC, or AaHapE deletion failed to incite necrotic lesions, likely due to severe growth deficiency. Our results revealed that siderophore biosynthesis and iron homeostasis are regulated by a well-organized network in A. alternata.


Introduction
Iron, often functioning as a cofactor of enzymes, is required for a variety of cellular processes, including deoxyribonucleotide synthesis, oxidative phosphorylation, activities of several enzymes, and detoxification of hydroxyl radicals generated from the Haber-Weiss/Fenton pathway [1,2]. However, excess iron in cells could cause cytotoxicity. Despite the abundance of total iron in the environment, most iron sources are insoluble ferric oxyhy-droxides, which are inaccessible to microorganisms [3]. It is estimated that less than 10−18 M iron in the environment is bioavailable to microorganisms [4]. Thus, microorganisms must have sophisticated mechanisms to acquire iron effectively from their environment.
Fungi have evolved three strategies: acidification of the environment, reduction of ferric iron (Fe 3+ ) to a more soluble ferrous form (Fe 2+ ) via the reductive iron assimilation (RIA) systems, and secretion of soluble iron-chelators called siderophores to acquire environmental iron [5,6]. The siderophore-Fe 3+ complex is delivered into cells via specific transporters [7]. In fungal cells, ferric iron can be stored in the vacuoles and as the siderophore-Fe 3+ complex in the cytoplasm [5,8]. The siderophore-mediated iron acquisition has been shown to be required for virulence in many pathogenic fungi of humans and plants [9][10][11].
Siderophore biosynthesis and iron acquisition must be tightly regulated to maintain iron homeostasis and avoid iron toxicity. Saccharomyces cerevisiae utilizes two transcription factors, Aft1 and Aft2, to regulate iron uptake, recycling, and mobilization [14][15][16]. Filamentous fungi utilize a GATA repressor (SreA) and a basic leucine zipper (bZIP)containing regulator (HapX) to control the expression of iron-acquisition genes and to maintain siderophore production and iron homeostasis [17,18]. HapX has been shown to interact with HapB and be recruited to bind to a CCAAT-binding complex (CBC) composed of three proteins, HapB, HapC, and HapE, in Aspergillus nidulans [19,20]. Under iron-depleted conditions, HapX promotes siderophore production and iron acquisition by suppressing the expression of SreA [17]. HapX is also known to suppress the genes involved in iron consumption, including heme cytochrome C, iron-sulfur cluster-containing aconitase, and homoaconitase under iron-depleted conditions [21][22][23][24]. In contrast, under iron-replete conditions, SreA suppresses the expression of HapX and many genes involved in siderophore biosynthesis and iron uptake [25]. Although both SreA and HapX are required to regulate siderophore biosynthesis, their roles in fungal virulence vary in different fungal species. SreA is not required for virulence in the human pathogens fumigatus [26] and Candida albicans [27] but is required for full virulence in the human pathogen Cryptococcus neoformans [28] and the maize pathogen C. heterostrophus [29]. HapX is required for virulence in A. fumigatus, C. albicans, and C. neoformans [22,27,30], as well as the plant pathogenic fungi Verticillium dahlia [24] and F. oxysporum [23].
The tangerine pathotype of A. alternata produces a host-selective toxin called ACT to kill susceptible citrus cultivars before colonization and acquire nutrients from dead tissues [31]. ACT is the primary pathogenicity determinant as the ∆AaACTT6 mutation strain fails to accumulate ACT and induce necrotic lesions in the susceptible citrus cultivar [32]. In addition, the detoxification of toxic reactive oxygen species (ROS) and the production of cutinases and siderophores are required for virulence [10,32,33]. During the invasion, A. alternata must overcome the toxicity of host-generated ROS and acquire iron from host tissues to establish successful colonization. Iron can facilitate the detoxification of hydrogen peroxide (H 2 O 2 ), promote the activity of antioxidant enzymes (e.g., catalase and superoxide dismutase), and assist fungal invasion [34]. Deleting an AaNps6 gene in the tangerine pathotype of A. alternata results in a mutant that fails to synthesize siderophores, increases sensitivity to H 2 O 2 , and reduces virulence in citrus [10], indicating that siderophore-mediated iron uptake plays a critical role in A. alternata pathogenesis.
A. alternata produces coprogens and hydroxycoprogens [35]. We have previously demonstrated that the biosynthesis of siderophores is negatively regulated by AaSreA under iron-replete conditions [36]. However, the AaSreA-deficient mutant displays wildtype virulence in citrus. To better understand the regulatory pathways controlling the biosynthesis of siderophores and iron uptake, we investigated the functions of AaHapX and three CCAAT-binding proteins (AaHapB, AaHapC, and AaHapE) in siderophore production, stress resistance, and virulence in A. alternata. Moreover, a yeast-activating transcription factor 1 (Atf1) homolog containing a bZIP domain was found to play a role in siderophore biosynthesis primarily by activating the expression of AaNps6. AaAtf1 was misidentified from the genome of A. alternata when searching for the yeast Aft1 and Aft2 homologs. Transcriptional interactions among these regulators under iron-depleted and replete conditions were established. This study demonstrated a complex regulatory network, directly and indirectly controlling siderophore biosynthesis and iron acquisition in the phytopathogenic fungus A. alternata.

Fungal Strains and Growth Conditions
The wild-type EV-MIL-31 strain of A. alternata was isolated from a diseased leaf of Minneola tangelo (Citrus x tangelo J.W. Ingram & H.E. Moore) and used in this study as a recipient host for transformation and the mutagenesis experiments. The AaSreA-deficient mutants (∆AaSreA_D6 and D12) carrying a deletion of the gene encoding a GATA-type transcription suppressor were obtained from a separate study [36]. ∆AaNps6, a defective strain for the biosynthesis of siderophores, was generated by Chen et al. (2013) [10]. ∆AaACTT6 carrying a mutation of the ACT biosynthetic gene, AaACTT6, was generated by Ma et al. (2019) [32]. Unless otherwise stated, fungal strains were grown on potato dextrose agar (PDA; Difco, Sparks, MD, USA) or minimal medium (MM) [36] at 28 • C for 3 to 5 days. For the isolation of nucleic acids, fungi were grown in a complete medium (MM+ 1 g/L yeast extract and 1 g/L casein hydrolysate). For ACT toxin production, fungi were grown in a modified Richard's medium [37] for 21 days. A regeneration medium (RMM) [38] was used to recover fungal transformants. For medium shift experiments, fungal strains were grown in PDB for 2 to 3 days. Mycelium was collected through a sterile filter paper, washed with sterile water, transferred to MM or MM amended with an appropriate compound, and incubated for an additional 24 h.

Targeted Gene Deletion and Genetic Complementation in A. alternata
A split marker-mediated transformation [38] was used to delete AaHapX, AaHapB, AaHapC, AaHapE, and AaAtf1 in A. alternata. A hygromycin phosphotransferase gene cassette (HYG, 2.5 kb) was split into two overlapping fragments HYg and hYG, using PCR with primer pairs M13F/hyg4 and hyg3/M13R, respectively, from an HYG-containing plasmid. Two split marker fragments, HYg::5 target gene DNA fragment and hYG::3 target gene fragment, were generated by two-step fusion PCR with gene-specific primers (Supplementary Materials Figures S1-S5) and directly transformed into protoplasts prepared from wild type using polyethylene glycol and CaCl 2 as previously described [39]. Fungal transformants were recovered from RMM amended with 200 µg/mL hygromycin (Roche Applied Science, Indianapolis, IN, USA), examined by PCR with a gene-specific primer pairing with a HYG primer (hyg3 or hyg4), and verified further by Southern blot analyses using gene-specific and HYG probes or PCR-restriction enzyme length polymorphism (PCR-RFLP) patterns (Supplementary Materials Figures S1-S5). Two independent mutants from each gene deletion were used for phenotypic analyses. Defective phenotypes were restored by transforming a functional copy of the gene into a deletion mutant. Each gene, including its endogenous promoter, was amplified by PCR with gene-specific primers and cloned into pCB1532 carrying a sulfonylurea-resistance (Sur) gene [40]. To facilitate cloning, recognition sequences of restriction endonucleases were incorporated into the primers. Fungal transformants were selected on RMM amended with 5 g/mL sulfonylurea (Chem Service, West Chester, PA, USA) and tested for restoration of radial growth on PDA or siderophore production. Oligonucleotide primers used for PCR and Southern blot analyses are listed in Supplementary Materials Table S1.

Miscellaneous Procedures for Manipulation of Nucleic Acids
Fungal genomic DNA was obtained using a Genomic DNA Mini Purification kit (BioKit, Taipei, Taiwan) or a phenol/chloroform DNA extraction protocol. Fungal RNA was isolated using TRI reagent (Sigma-Aldrich, St. Louis, MO, USA) and purified further using PureLink RNA Mini Kit (Invitrogen, Carlsbad, CA, USA) following the manufacturer's instructions. DNA digestion with restriction endonucleases, electrophoresis, DNA ligation, bacterial transformation, Southern blot hybridizations, and post-hybridization washing were carried out according to standard procedures. Digoxigenin (DIG)-11-dUTP (Roche Applied Science, Indianapolis, IN, USA) was incorporated by PCR into a DNA fragment with the gene-specific primers and used for Southern blot analyses. The probe was detected by an immunological assay according to the manufacturer's instructions (Roche Applied Science, Indianapolis, IN, USA). The protein-coding genes were predicted by GlimmerHMM [41,42]. Pairwise sequence comparison was performed using CLC Genomic Workbench 9.5.1 (CLC Bio, Qiagen, Aarhus, Denmark) to calculate genetic distance using the Jukes-Cantor model and percent identity. Conserved domains were predicted with the CD-search tool [43] available in the National Center for Biotechnology Information (NCBI) and the MEME Suite server [44].

Assays for Siderophore Production
The production of siderophores was assayed on a Chrome azurol S (CAS)-containing medium as described [45]. The formation of orange halos on the blue background around the fungal colony indicated the production of siderophores. The plates were photographed, and the radius of orange halos was determined using Image J software V. 1.54b (US National Institutes of Health, Bethesda, MD, USA). Siderophores were further isolated from culture filtrates of fungal strains cultured in MM for 5 to 7 days with Amberlite XAD-16 resin (Sigma-Aldrich, St. Louis, MO, USA) and examined by thin-layer chromatography (TLC) and high-performance liquid chromatography (HPLC) using mobile phases and conditions as described [35].

Quantitative RT-PCR and Gene Expression Analyses
Gene expression was evaluated by quantitative real-time PCR (qRT-PCR) using genespecific primers (Supplementary Materials Table S1). All qRT-PCR reactions were set up using iQ SYBR Green Supermix (Bio-Rad, Hercules, CA, USA) and performed in a CFX Connect model of Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA). Amplification of the β-tubulin coding gene was included as an internal control, and the specific transcripts were assessed by the melting curve. The relative expression level was determined by a comparative Ct method (∆∆CT). All treatments were conducted with three biological replicates, and the significant difference was determined by statistical analysis. Experiments were repeated at least two times.

Assays for Virulence and Toxin Production
A. alternata virulence was assessed on detached leaves of calamondin, a cross between a mandarin orange (Citrus reticulata Blanco) and a kumquat (Fortunella margarita Swingle). Leaves (6 to 8 days after emergence and approximately 3 to 4 cm in length) were harvested, washed with water, and inoculated with agar plugs carrying fungal hyphae. Some leaf spots were wounded by a fine needle before inoculation. Alternatively, leaves were inoculated by placing 5 µL of conidia suspensions (2 × 10 5 conidia/mL) on each spot. Leaves treated with water were used as mock controls. All leaves were kept in a plastic container for 3 to 5 days. ACT was purified with Amberlite XAD-2 resin (Sigma-Aldrich, St. Louis, MO, USA), separated in a silica gel 60 F254 plate (Merck KGaA, Darmstadt, Germany) using methanol/acetic acid/water (4:1:5, v/v) as mobile phase, visualized under a hand-held UV illuminator, and photographed.

Statistical Analysis
Unless stated otherwise, all experiments with multiple replicates were conducted at least two times. The significance of treatments was analyzed by one-way ANOVA and separated by Tukey's HSD post hoc test (p < 0.05).

Statistical Analysis
Unless stated otherwise, all experiments with multiple replicates were conducted at least two times. The significance of treatments was analyzed by one-way ANOVA and separated by Tukey's HSD post hoc test (p < 0.05).

Transcription Regulators Are Required for Growth and Iron Homeostasis
Two split HYG marker gene fragments were generated from each of the five genes and used for targeted gene deletion in the wild-type strain of A. alternata. As a result, at least two mutants were identified from each gene after verification by Southern blot analyses or RFLP patterns. Only mutants carrying a single HYG insertion were selected for further analyses. The AaHapX deficient mutants (∆AaHapX_X6 and X10) produced fewer aerial hyphae than the wild type on PDA and MM ( Figure 2). Adding BPS (an iron chelator) into PDA or MM reduced ∆AaHapX growth considerably. Adding FeCl 3 into MM but not PDA also reduced ∆AaHapX growth. The growth of fungal strains carrying gene deletion in AaHapB (∆AaHapB_B7 and B8), AaHapC (∆AaHapC_C12 and C45), or AaHapE (∆AaHapE_E1 and E2) was severely impaired, especially on MM amended with BPS ( Figure 2). Compared to the wild type, AaAtf1 deficient mutants (∆AaAtf1_A2 and A3) reduced growth on PDA by 31%. In contrast, ∆AaAtf1 grew slightly better than the wild type on MM. Adding BPS or FeCl 3 into PDA or MM did not affect ∆AaAtf1 growth. Compared to the wild type, the AaSreA deficient mutants reduced growth on PDA and MM by 80% and 40%, respectively. Adding BPS into PDA greatly enhanced ∆AaSreA growth, whereas adding FeCl 3 into MM reduced ∆AaSreA growth. Defective growth in each gene mutation was restored to the wild-type level after re-expressing a functional copy of the gene in the respective mutant ( Figure 2). J. Fungi 2023, 9, x FOR PEER REVIEW 6 of 17 transcription factors. Three cysteine-rich regions conserved in the AaHapX are also indicated. AaHapB, AaHapC, and AaHapE are CCAAT-binding proteins.

Transcription Regulators Are Required for Growth and Iron Homeostasis
Two split HYG marker gene fragments were generated from each of the five genes and used for targeted gene deletion in the wild-type strain of A. alternata. As a result, at least two mutants were identified from each gene after verification by Southern blot analyses or RFLP patterns. Only mutants carrying a single HYG insertion were selected for further analyses. The AaHapX deficient mutants (∆AaHapX_X6 and X10) produced fewer aerial hyphae than the wild type on PDA and MM ( Figure 2). Adding BPS (an iron chelator) into PDA or MM reduced ∆AaHapX growth considerably. Adding FeCl 3 into MM but not PDA also reduced ∆AaHapX growth. The growth of fungal strains carrying gene deletion in AaHapB (∆AaHapB_B7 and B8), AaHapC (∆AaHapC_C12 and C45), or AaHapE (∆AaHapE_E1 and E2) was severely impaired, especially on MM amended with BPS ( Figure 2). Compared to the wild type, AaAtf1 deficient mutants (∆AaAtf1_A2 and A3) reduced growth on PDA by 31%. In contrast, ∆AaAtf1 grew slightly better than the wild type on MM. Adding BPS or FeCl 3 into PDA or MM did not affect ∆AaAtf1 growth. Compared to the wild type, the AaSreA deficient mutants reduced growth on PDA and MM by 80% and 40%, respectively. Adding BPS into PDA greatly enhanced ∆AaSreA growth, whereas adding FeCl 3 into MM reduced ∆AaSreA growth. Defective growth in each gene mutation was restored to the wild-type level after re-expressing a functional copy of the gene in the respective mutant ( Figure 2).

Figure 2.
Transcription regulators involved in iron homeostasis and growth in A. alternata under iron-replete (PDA) and iron-depleted (MM) conditions. Fungal strains, including wild type (WT), deletion mutants (∆), and complementation strains (CP) were transferred by pipette tips to PDA and MM amended with or without 0.2 mM FeCl 3 or 0.2 mM bathophenanthrolinedisulfonic acid (BPS) and incubated at 28 °C for 3 to 5 days. ∆AaHapX grew poorly on a medium amended with BPS (indicated by yellow rectangles). However, ∆AaSreA grew poorly under iron-rich conditions (indicated by a red rectangle), and exhibited wild-type growth in the presence of BPS (indicated by a green rectangle). ∆AaHapB, ∆AaHapC, and ∆AaHapE grew poorly on either medium. Experiments were tested at least two times with two replicates for each treatment. Transcription regulators involved in iron homeostasis and growth in A. alternata under iron-replete (PDA) and iron-depleted (MM) conditions. Fungal strains, including wild type (WT), deletion mutants (∆), and complementation strains (CP) were transferred by pipette tips to PDA and MM amended with or without 0.2 mM FeCl 3 or 0.2 mM bathophenanthrolinedisulfonic acid (BPS) and incubated at 28 • C for 3 to 5 days. ∆AaHapX grew poorly on a medium amended with BPS (indicated by yellow rectangles). However, ∆AaSreA grew poorly under iron-rich conditions (indicated by a red rectangle), and exhibited wild-type growth in the presence of BPS (indicated by a green rectangle). ∆AaHapB, ∆AaHapC, and ∆AaHapE grew poorly on either medium. Experiments were tested at least two times with two replicates for each treatment.

Biosynthesis of Siderophores Is Positively and Negatively Regulated by Multiple Transcription Factors
CAS and TLC assays of siderophores revealed that the production of siderophores was severely impaired in the ∆AaHapX, ∆AaHapB, ∆AaHapC, ∆AaHapE, and ∆AaAtf1 strains ( Figure 3). In contrast, the production of siderophores was greatly increased in ∆AaSreA. The complementation strains in each of the gene mutations had the wild-type level of siderophore production. Siderophores produced by the wild-type, the mutants (∆AaHapX and ∆AaAtf1), and their complementation strains were also confirmed by HPLC analyses (Supplementary Materials Figure S11).

Biosynthesis of Siderophores Is Positively and Negatively Regulated by Multiple Transcription Factors
CAS and TLC assays of siderophores revealed that the production of siderophores was severely impaired in the ∆AaHapX, ∆AaHapB, ∆AaHapC, ∆AaHapE, and ∆AaAtf1 strains ( Figure 3). In contrast, the production of siderophores was greatly increased in ∆AaSreA. The complementation strains in each of the gene mutations had the wild-type level of siderophore production. Siderophores produced by the wild-type, the mutants (∆AaHapX and ∆AaAtf1), and their complementation strains were also confirmed by HPLC analyses (Supplementary Materials Figure S11).

Cross-Interactions between Transcriptional Regulators under Different Iron Conditions
Quantitative RT-PCR analyses revealed that AaHapX was preferably expressed under iron-depleted conditions, whereas AaSreA was highly expressed under iron-replete conditions ( Figure 4). Iron had little or no effect on the expression of AaAtf1, AaHapB, AaHapC, and AaHapE (Figures 4 and S12). Deleting AaHapX increased the expression of

Cross-Interactions between Transcriptional Regulators under Different Iron Conditions
Quantitative RT-PCR analyses revealed that AaHapX was preferably expressed under iron-depleted conditions, whereas AaSreA was highly expressed under iron-replete conditions ( Figure 4). Iron had little or no effect on the expression of AaAtf1, AaHapB, AaHapC, and AaHapE (Figures 4 and S12). Deleting AaHapX increased the expression of AaSreA under iron-depleted conditions. Under iron-rich conditions, deleting AaHapX had little or no effect on the expression of AaSreA. Deleting AaHapX slightly increased the expression of AaAtf1, particularly under iron-depleted conditions. Deleting AaSreA significantly 245 increased the expression of AaHapX and AaAtf1, primarily under iron-replete conditions. Deleting AaAtf1 had little or no effect on the expression of AaHapX and AaSreA under either iron-depleted or replete conditions. AaSreA under iron-depleted conditions. Under iron-rich conditions, deleting AaHapX ha little or no effect on the expression of AaSreA. Deleting AaHapX slightly increased th expression of AaAtf1, particularly under iron-depleted conditions. Deleting AaSreA significantly 245 increased the expression of AaHapX and AaAtf1, primarily under iron replete conditions. Deleting AaAtf1 had little or no effect on the expression of AaHapX an AaSreA under either iron-depleted or replete conditions. , and ∆AaAtf1 (A2 and A3) were grown in PDB at 2 °C for 2 to 3 days. Mycelium was harvested, washed with water, transferred into liquid MM or MM amended with 0.2 mM FeCl 3 , and incubated for an additional 24 h. RNA was purified from mycelium and used for first-strand cDNA synthesis. Quantitative RT-PCR was performed using gene-specif primers. The relative expression level from three independent reactions was calculated by comparative Ct method (∆∆CT) in relation to the expression of the fungal β-tubulin-coding gene. I each assay, the expression level in the wild type grown in MM was set to 1. Means indicated by th same letter are not significantly different, p < 0.05. All experiments were repeated at least two time showing similar trends.

The Genes Involved in Siderophore Biosynthesis and Iron Acquisition Are Differentially Regulated by AaHapX and AaAtf1
Under iron-depleted conditions, deleting AaHapX decreased the expression of AaNps (encoding a nonribosomal peptide synthetase) ( Figure 5). In contrast, under iron-replet conditions, deleting AaHapX increased the expression of AaNps6, AaSit1 (encoding siderophore iron transporter), and AaMirB (encoding a siderophore iron transporter Deleting AaAtf1 decreased the expression of AaNps6 under iron-depleted conditions however, this had no effect on the expression of AaNps6, AaSit1, and AaMirB under iron replete conditions. and AaAtf1 (G-I) genes in A. alternata strains grown in the presence or absence of iron. Wild type, ∆AaHapX (X6 and X10), ∆AaSreA (D6 and D12), and ∆AaAtf1 (A2 and A3) were grown in PDB at 28 • C for 2 to 3 days. Mycelium was harvested, washed with water, transferred into liquid MM or MM amended with 0.2 mM FeCl 3 , and incubated for an additional 24 h. RNA was purified from mycelium and used for first-strand cDNA synthesis. Quantitative RT-PCR was performed using gene-specific primers. The relative expression level from three independent reactions was calculated by a comparative Ct method (∆∆CT) in relation to the expression of the fungal β-tubulin-coding gene. In each assay, the expression level in the wild type grown in MM was set to 1. Means indicated by the same letter are not significantly different, p < 0.05. All experiments were repeated at least two times, showing similar trends.

The Genes Involved in Siderophore Biosynthesis and Iron Acquisition Are Differentially Regulated by AaHapX and AaAtf1
Under iron-depleted conditions, deleting AaHapX decreased the expression of AaNps6 (encoding a nonribosomal peptide synthetase) ( Figure 5). In contrast, under iron-replete conditions, deleting AaHapX increased the expression of AaNps6, AaSit1 (encoding a siderophore iron transporter), and AaMirB (encoding a siderophore iron transporter). Deleting AaAtf1 decreased the expression of AaNps6 under iron-depleted conditions; however, this had no effect on the expression of AaNps6, AaSit1, and AaMirB under iron-replete conditions. Figure 5. Quantitative real-time PCR analyses of the expression of siderophore biosynthesis-and transport-related genes in A. alternata strains grown in the presence or absence of iron. Wild type, ∆AaHapX (X6 and X10), and ∆AaAtf1 (A2 and A3) were grown in PDB at 28 °C for 2 to 3 days. Mycelium was harvested, washed with water, transferred into liquid MM or MM amended with 0.2 mM FeCl 3 , and incubated for an additional 24 h. RNA was purified from mycelium and used for first-strand cDNA synthesis. Quantitative RT-PCR was performed using primers targeting AaNps6 (A,B), AaSit1 (C,D), or AaMirB (E,F). The relative expression level from three independent reactions was calculated by a comparative Ct method (∆∆CT) in relation to the expression of the fungal βtubulin-coding gene. In each assay, the expression level in the wild type grown in MM was set to 1. Means indicated by the same letter are not significantly different, p < 0.05. All experiments were repeated at least two times, showing similar trends.

AaAtf1 and AaHapX Play a Negative Role in Osmotic Stress under Nutrient-Rich Conditions
Sensitivity tests assayed on PDA revealed that deleting AaAtf1 increased resistance to high glucose, sucrose, mannitol, and sorbitol but not NaCl and KCl ( Figure 6A). Compared to the wild type, ∆AaAtf1 reduced radial growth by 31% on PDA, whereas ∆AaAtf1 exhibited growth similar to wild type on PDA amended with a sugar osmoticant. In contrast, deleting AaHapX increased resistance to NaCl and KCl but not sugar osmoticants ( Figure 6B). ∆AaAtf1 showed no significant difference in growth inhibition on MM amended with or without osmoticants (Supplementary Materials Figure S13). When assayed on PDA and MM, deleting AaAtf1 did not affect resistance or sensitivity to H 2 O 2 , tBHP, menadione, diethyl maleate, and Congo red. Deleting AaHapX also unchanged sensitivity to H 2 O 2 (Supplementary Materials Figure S13). Mycelium was harvested, washed with water, transferred into liquid MM or MM amended with 0.2 mM FeCl 3 , and incubated for an additional 24 h. RNA was purified from mycelium and used for first-strand cDNA synthesis. Quantitative RT-PCR was performed using primers targeting AaNps6 (A,B), AaSit1 (C,D), or AaMirB (E,F). The relative expression level from three independent reactions was calculated by a comparative Ct method (∆∆CT) in relation to the expression of the fungal β-tubulin-coding gene. In each assay, the expression level in the wild type grown in MM was set to 1. Means indicated by the same letter are not significantly different, p < 0.05. All experiments were repeated at least two times, showing similar trends.

AaAtf1 and AaHapX Play a Negative Role in Osmotic Stress under Nutrient-Rich Conditions
Sensitivity tests assayed on PDA revealed that deleting AaAtf1 increased resistance to high glucose, sucrose, mannitol, and sorbitol but not NaCl and KCl ( Figure 6A). Compared to the wild type, ∆AaAtf1 reduced radial growth by 31% on PDA, whereas ∆AaAtf1 exhibited growth similar to wild type on PDA amended with a sugar osmoticant. In contrast, deleting AaHapX increased resistance to NaCl and KCl but not sugar osmoticants ( Figure 6B). ∆AaAtf1 showed no significant difference in growth inhibition on MM amended with or without osmoticants (Supplementary Materials Figure S13). When assayed on PDA and MM, deleting AaAtf1 did not affect resistance or sensitivity to H 2 O 2 , tBHP, menadione, diethyl maleate, and Congo red. Deleting AaHapX also unchanged sensitivity to H 2 O 2 (Supplementary Materials Figure S13).

Siderophore-Related Regulators Play No Role in Toxin Production and A. alternata Virulence
Pathogenicity assays were conducted on detached calamondin leaves, revealing that ∆AaAtf1 and ∆AaHapX induced necrotic lesions at rates and magnitudes similar to the wild type 3 days post inoculation (dpi) ( Figure 7A). Because ∆AaHapB, ∆AaHapC, and ∆AaHapE grew slowly and produced few conidia, their pathogenicity was assessed on detached leaves using mycelial mass. The results revealed that ∆AaHapB, ∆AaHapC, and ∆AaHapE were unable to induce necrotic lesions 3 dpi, even when the leaves were wounded before inoculation. ACT was isolated from culture filtrates of the wild-type, the ∆AaAtf1, and the CPA1 complementation strains and analyzed by TLC, revealing no significant difference in the level of ACT ( Figure 7B). As mock controls, no ACT was detected from medium-only or culture filtrates of an ACT deficient strain (∆AaACTT6).

Siderophore-Related Regulators Play No Role in Toxin Production and A. alternata Virulence
Pathogenicity assays were conducted on detached calamondin leaves, revealing that ∆AaAtf1 and ∆AaHapX induced necrotic lesions at rates and magnitudes similar to the wild type 3 days post inoculation (dpi) ( Figure 7A). Because ∆AaHapB, ∆AaHapC, and ∆AaHapE grew slowly and produced few conidia, their pathogenicity was assessed on detached leaves using mycelial mass. The results revealed that ∆AaHapB, ∆AaHapC, and ∆AaHapE were unable to induce necrotic lesions 3 dpi, even when the leaves were wounded before inoculation. ACT was isolated from culture filtrates of the wild-type, the ∆AaAtf1, and the CPA1 complementation strains and analyzed by TLC, revealing no significant difference in the level of ACT ( Figure 7B). As mock controls, no ACT was detected from medium-only or culture filtrates of an ACT deficient strain (∆AaACTT6).

Figure 7.
AaHapX and AaAtf1 play no role in A. alternata pathogenesis. (A) Pathogenicity tests assayed on detached calamondin leaves revealed that ∆AaHapX (X6 and X10) and ∆AaAtf1 (A2 and A3) induced necrotic lesions at rates and magnitudes similar to the wild type and their respective complementation (CP) strains in pre-wounded (W) or un-wounded (Un) leaves. In contrast, ∆AaHapB (B7 and B8), ∆AaHapC (C12 and C45), and ∆AaHapE (E1 and E2) barely induced necrotic lesions. (B) TLC analysis of toxin purified from wild type, ∆AaAtf1 (A2 and A3), and the complementation strain (CPA1) revealed that deleting AaAtf1 had no effects on ACT production. No ACT was detected from medium only or culture filtrates of ∆AaACTT6 (an ACT deficient strain). The arrow indicates the direction of the mobile phase (solvent).

Discussion
Iron is an essential trace element required for metabolic, enzymatic, and regulatory functions in all living cells [46]. In contrast, excess iron could be toxic. Thus, all cells must have elaborate systems to maintain iron homeostasis [47,48]. Previously, we have demonstrated that, under iron-replete conditions, A. alternata utilizes AaSreA to suppress the production and transport of siderophores to avoid iron toxicity [36]. In the present study, we characterized five transcription regulators that positively controlled the biosynthesis of siderophores in the phytopathogenic fungus A. alternata. Our results indicated that alternata mainly relies on a transcriptional feedback inhibition between the AaSreA iron repressor and the AaHapX transcription factor to regulate iron acquisition (Figure 8). AaHapX mainly functions under iron-depleted conditions, whereas AaSreA functions under iron-replete conditions. (B) TLC analysis of toxin purified from wild type, ∆AaAtf1 (A2 and A3), and the complementation strain (CPA1) revealed that deleting AaAtf1 had no effects on ACT production. No ACT was detected from medium only or culture filtrates of ∆AaACTT6 (an ACT deficient strain). The arrow indicates the direction of the mobile phase (solvent).

Discussion
Iron is an essential trace element required for metabolic, enzymatic, and regulatory functions in all living cells [46]. In contrast, excess iron could be toxic. Thus, all cells must have elaborate systems to maintain iron homeostasis [47,48]. Previously, we have demonstrated that, under iron-replete conditions, A. alternata utilizes AaSreA to suppress the production and transport of siderophores to avoid iron toxicity [36]. In the present study, we characterized five transcription regulators that positively controlled the biosynthesis of siderophores in the phytopathogenic fungus A. alternata. Our results indicated that A. alternata mainly relies on a transcriptional feedback inhibition between the AaSreA iron repressor and the AaHapX transcription factor to regulate iron acquisition (Figure 8). AaHapX mainly functions under iron-depleted conditions, whereas AaSreA functions under iron-replete conditions. AaHapX was recruited to bind to the AaHapB/AaHapC/AaHapE complex, which in turn suppressed the expression of AaSreA encoding a repressor against siderophore biosynthesis. Without AaSreA, the genes involved in siderophore biosynthesis, transport, and iron uptake were de-repressed. As such, siderophores were produced to chelate iron from the environment. AaAtf1 enhanced the expression of AaNps6, thus playing a positive role in siderophore production. Under high-iron conditions, the expression of AaSreA was upregulated, which in turn suppressed the expression of AaHapX, AaAtf1, and genes involved in siderophore biosynthesis, thus preventing siderophore production.
The A. alternata SreA encodes a polypeptide containing two conserved GATA-type zinc finger domains and plays a determinant role in the biosynthesis of siderophores [36]. The expression of AaSreA was upregulated under iron-rich conditions and downregulated under iron-depleted conditions. In contrast, the expression of AaHapX was downregulated under iron-rich conditions and upregulated under iron-depleted conditions. Thus, AaSreA serves as a repressor under iron-replete conditions, and the bZIP transcription regulator AaHapX acts as an activator under iron-depleted conditions. When environmental iron is scarce, the expression of AaSreA is inhibited by AaHapX (Figure 8). AaHapX may also directly upregulates the gene (AaNps6) involved in the biosynthesis of siderophores under iron-depleted conditions. When environmental iron is abundant, AaSreA is highly expressed, and the AaSreA protein suppresses the expression of the genes involved in the biosynthesis and transportation of siderophores. AaSreA also suppressed the expression of the AaHapX gene under iron-rich conditions. The results indicated that AaSreA and AaHapX are the major siderophore regulators by forming a negative feedback loop to 305 regulate iron acquisition in response to the availability of environmental iron in A. alternata. A similar transcriptional feedback inhibition between SreA and HapX has been reported in Aspergillus nidulans [17], confirming the important role of SreA and HapX in siderophore production in filamentous fungi.
HapX interacts with the CCAAT binding proteins (HapB, HapC, and HapE) and forms a CBC complex involved in suppressing the expression of SreA and activating the biosynthesis of siderophores under iron-depleted conditions in Aspergillus spp. [17,47,49]. The function of HapX is highly dependent on the CBC/HapX interaction [50]; however, deleting the HapX-coding gene homolog could lead to different deficiencies in different fungal species. In A. alternata, deleting AaHapX exhibited reduced siderophore AaHapX was recruited to bind to the AaHapB/AaHapC/AaHapE complex, which in turn suppressed the expression of AaSreA encoding a repressor against siderophore biosynthesis. Without AaSreA, the genes involved in siderophore biosynthesis, transport, and iron uptake were de-repressed. As such, siderophores were produced to chelate iron from the environment. AaAtf1 enhanced the expression of AaNps6, thus playing a positive role in siderophore production. Under high-iron conditions, the expression of AaSreA was upregulated, which in turn suppressed the expression of AaHapX, AaAtf1, and genes involved in siderophore biosynthesis, thus preventing siderophore production.
The A. alternata SreA encodes a polypeptide containing two conserved GATA-type zinc finger domains and plays a determinant role in the biosynthesis of siderophores [36]. The expression of AaSreA was upregulated under iron-rich conditions and downregulated under iron-depleted conditions. In contrast, the expression of AaHapX was downregulated under iron-rich conditions and upregulated under iron-depleted conditions. Thus, AaSreA serves as a repressor under iron-replete conditions, and the bZIP transcription regulator AaHapX acts as an activator under iron-depleted conditions. When environmental iron is scarce, the expression of AaSreA is inhibited by AaHapX (Figure 8). AaHapX may also directly upregulates the gene (AaNps6) involved in the biosynthesis of siderophores under irondepleted conditions. When environmental iron is abundant, AaSreA is highly expressed, and the AaSreA protein suppresses the expression of the genes involved in the biosynthesis and transportation of siderophores. AaSreA also suppressed the expression of the AaHapX gene under iron-rich conditions. The results indicated that AaSreA and AaHapX are the major siderophore regulators by forming a negative feedback loop to 305 regulate iron acquisition in response to the availability of environmental iron in A. alternata. A similar transcriptional feedback inhibition between SreA and HapX has been reported in Aspergillus nidulans [17], confirming the important role of SreA and HapX in siderophore production in filamentous fungi.
HapX interacts with the CCAAT binding proteins (HapB, HapC, and HapE) and forms a CBC complex involved in suppressing the expression of SreA and activating the biosynthesis of siderophores under iron-depleted conditions in Aspergillus spp. [17,47,49]. The function of HapX is highly dependent on the CBC/HapX interaction [50]; however, deleting the HapX-coding gene homolog could lead to different deficiencies in different fungal species. In A. alternata, deleting AaHapX exhibited reduced siderophore production; however, this had a moderate impact on growth, whereas deleting AaHapB, AaHapC, or AaHapE resulted in severe growth retardation, indicating that AaHapB/C/E might also be independent of AaHapX to regulate the expression of genes required for growth in A. alternata. The results also suggest that AaHapX primarily responds to iron availability. In F. oxysporum, deleting the HapX homolog upregulates the genes involved in the ironconsuming pathways but has no effect on the iron acquisition, resulting in growth reduction under iron-depleted conditions and lower virulence in hosts [23]. In V. dahlia, HapX is required for iron homeostasis, growth, the formation of conidia and microsclerotia, resistance to H 2 O 2 , and virulence [24]. Sensitivity assays revealed that AaHapX plays no role in resistance to oxidative stress and iron resistance in A. alternata. The function of HapX in regulating siderophore biosynthesis and virulence has also been reported in the opportunistic pathogens A. fumigatus, C. albicans, and C. neoformans of humans [22,27,30] and in the insect mycopathogen Beauveria bassiana [51]. In contrast, deleting AaHapX in A. alternata had no impact on virulence, even though the mutant is defective in siderophore production. Deleting the AaSreA repressor-coding gene resulting in high siderophore production also has no impact on A. alternata virulence [36]. However, deleting the AaNps6 gene completely blocks siderophore biosynthesis and results in lower virulence in citrus [10]. Thus, a low-level production of siderophores might be sufficient to enable A. alternata to colonize citrus hosts as long as toxin production is unaffected, as was revealed in the ∆AaAtf1 mutant. ∆AaHapB, ∆AaHapC, and ∆AaHapE failed to produce necrotic lesions mainly due to severe growth defects. In addition to siderophore production, AaHapX confers sensitivity to salts in A. alternata as the ∆AaHapX mutant grew much better than the wild type on PDA amended with NaCl or KCl. This is a novel phenotype associated with AaHapX. Interconnection between iron homeostasis and osmotic stress response has also been reported in plants [52] and the halophilic bacterium, Chromohalobacter salexigens [53].
In addition to AaSreA and AaHapX/B/C/E, we found that AaAtf1 containing a bZIP_ATF2 domain at its C terminus is involved in siderophore biosynthesis. This novel phenotype regarding Atf1 and siderophore production has never been reported in any fungi. Atf1 has been shown to be involved in stress response, sexual development, and meiotic hot spot activation in the fission yeast Schizosaccharomyces pombe [54,55], Cryptococcus neoformans [56] and many filamentous fungi, including A. nidulans [57], A. oryzae [58], N. crassa [59], A. fumigatus [57], Mucor circinelloides [60], Penicillium marneffei [61], F. graminearum [62], F. verticillioides [63], F. oxysporum f. sp. cubense [64], V. dahlia [65], Claviceps purpurea [66], Botrytis cinerea [67], and M. oryzae [68]. Depending on fungal species, Atf1 plays a role in resistance to heat, cold, desiccation, cell wall-disrupting agents (Congo red and calcofluor white), fungicides (fludioxonil and caspofungin), osmotic or oxidative stress, and the maintenance of iron homeostasis [69]. AaAtf1 plays a negative role in osmotic resistance under nutrient-rich conditions, consistent with the finding that AaAtf1 contains an osmotic stress response (OSM) domain. However, AaAtf1 is not required for fungal virulence as assayed on detached citrus leaves. Sensitivity tests revealed that ∆AaAtf1 grew much slower than wild type on PDA and displayed wild-type growth on PDA but not MM amended with glucose, sucrose, mannitol, or sorbitol. Unlike other fungal species, AaAtf1 plays no role in resistance to cell-wall-disrupting agents, heat/cold, or oxidative stress. Surprisingly, AaAtf1 is involved in siderophore biosynthesis by regulating the expression of AaNps6. The results indicate that Atf1 homologs evolve in different fungi, providing a better adaptation to their biological environments.
In conclusion, siderophore-mediated iron uptake is required for growth, conidiation, ROS resistance, and virulence in A. alternata. This fungus depends on the AaSreA repressor and the AaHapX/CBC complex activator in response to the iron availability in the environment. Siderophore biosynthesis is also regulated by AaAtf1. AaAtf1 activates the expression of AaNps6 under iron-depleted conditions and facilitates siderophore biosynthesis. Although both AaAtf1 and AaHapX are involved in siderophore production, AaAtf1 plays a negative role in resistance to sugar-induced osmotic stress, and AaHapX plays a negative role in resistance to salt-induced osmotic stress under nutrient-rich conditions. This study further highlights that a well-regulated network is orchestrated to control the biosynthesis of siderophore, iron uptake, and stress response in A. alternata. Further studies using RNA sequencing analyses will allow us to identify key or novel genes that are regulated by AaHapX, AaAtf1, and AaSreA.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/jof9040427/s1, Figure S1: Deletion of the AaAtf1 gene using a split marker approach in the tangerine pathotype of A. alternata; Figure S2: Deletion of the AaHapX gene using a split marker approach in the tangerine pathotype of A. alternata; Figure S3: Deletion of the AaHapB gene using a split marker approach in the tangerine pathotype of A. alternata; Figure S4: Deletion of the AaHapC gene using a split marker approach in the tangerine pathotype of A. alternata; Figure S5: Deletion of the AaHapE gene using a split marker approach in the tangerine pathotype of A. alternata; Figure S6: Pairwise comparison and alignment of conserved motifs of the AaAtf1 sequence with fungal Atf1 orthologs; Figure S7: Pairwise comparison and alignment of conserved motifs of the AaHapX sequence with fungal HapX orthologs; Figure S8: Pairwise comparison and alignment of conserved motifs of the AaHapB sequence with fungal HapB orthologs; Figure S9: Pairwise comparison and alignment of conserved motifs of the AaHapC sequence with fungal HapC orthologs; Figure S10: Pairwise comparison and alignment of conserved motifs of the AaHapE sequence with fungal HapE orthologs; Figure S11: HPLC analyses of siderophores isolated from the strains of A. alternata; Figure S12: qRT-PCR analyses; Figure S13: Sensitivity tests; Table S1: Oligonucleotide primers.