The NADPH Oxidase A of Verticillium dahliae Is Essential for Pathogenicity, Normal Development, and Stress Tolerance, and It Interacts with Yap1 to Regulate Redox Homeostasis

Maintenance of redox homeostasis is vital for aerobic organisms and particularly relevant to plant pathogens. A balance is required between their endogenous ROS production, which is important for their development and pathogenicity, and host-derived oxidative stress. Endogenous ROS in fungi are generated by membrane-bound NADPH oxidase (NOX) complexes and the mitochondrial respiratory chain, while transcription factor Yap1 is a major regulator of the antioxidant response. Here, we investigated the roles of NoxA and Yap1 in fundamental biological processes of the important plant pathogen Verticillium dahliae. Deletion of noxA impaired growth and morphogenesis, compromised formation of hyphopodia, diminished penetration ability and pathogenicity, increased sensitivity against antifungal agents, and dysregulated expression of antioxidant genes. On the other hand, deletion of yap1 resulted in defects in conidial and microsclerotia formation, increased sensitivity against oxidative stress, and down-regulated antioxidant genes. Localized accumulation of ROS was observed before conidial fusion and during the heterokaryon incompatibility reaction upon nonself fusion. The frequency of inviable fusions was not affected by the deletion of Yap1. Analysis of a double knockout mutant revealed an epistatic relationship between noxA and yap1. Our results collectively reveal instrumental roles of NoxA and ROS homeostasis in the biology of V. dahliae.


Introduction
Reactive oxygen species (ROS) are short-lived, highly reactive molecules that are produced by partial reduction of oxygen, and they include hydrogen peroxide (H 2 O 2 ), superoxide anion (O 2 •− ) and hydroxyl (OH • ) radicals [1]. Growing evidence from various organisms has assigned a Janus-faced nature to ROS as both deleterious molecules that can cause irreversible damage to biological systems and important signaling components regulating metabolic and developmental processes [1][2][3]. This double-edged significance of ROS is clearly illustrated in host-microbe interactions. In particular, fungal plant pathogens have to cope with host-derived ROS burst reactions during infection [4], while endogenous ROS production is simultaneously involved in the development of infectious structures and the efficient colonization of their hosts [5]. It is, therefore, vital for fungal pathogens to fungi. To this end, we knocked out their V. dahliae homologs and further generated a double knockout mutant to investigate their roles in fungal development, physiology, and pathogenicity. Furthermore, we characterized the responses of these mutants to antifungal agents and their behavior during hyphal fusion and the heterokaryon incompatibility reaction. Our results attribute significant roles to NoxA and Yap1 regarding morphogenesis, pathogenicity, and stress tolerance, and they highlight the crucial contribution of endogenous ROS production and metabolism to multiple aspects of fungal biology.

Fungal Strains, Growth Media, and Culture Conditions
All fungal strains constructed and used in this study are listed in Table S1 (Supplementary Materials). They were grown on standard growth media (Potato Dextrose Agar-PDA, Czapek-Dox complete medium-CM, Czapek-Dox minimal medium-MM) at 24 • C, in the dark. Preparation and maintenance of monoconidial strains have been described previously [46].

Deletion and Complementation of V. dahliae noxA and yap1
All plasmids used for the construction of recombinant vectors are listed in Table S2 (Supplementary Materials). We have recently described the strategy that we used for knocking out the V. dahliae homologs of noxA and yap1 [23]. Briefly, both mutants were constructed in the background of V. dahliae wild-type strain 123V via double homologous recombination following Agrobacterium tumefaciens-mediated transformation (ATMT) [47]. Using this method, we recently performed and described deletion of noxA and validation of the mutant strain [23]. To knock out the yap1 homolog, the flanking chromosomal regions (~2.0 kb-long each) of the yap1 open reading frame (ORF) were amplified from genomic DNA of V. dahliae 123V and ligated to the neo R cassette (conferring resistance to geneticin; the cassette was amplified from plasmid pSD1 [48]) in the backbone of the A. tumefaciens binary vector pOSCAR [49]. All PCR amplification steps were performed using the high-fidelity Herculase II Fusion DNA Polymerase (Agilent, Santa Clara, CA, USA), and assembly of the recombinant vector was performed with the NEBuilder HiFi DNA Assembly Master Mix (New England Biolabs, Ipswich, MA, USA). To generate the double ∆noxA ∆yap1 knockout mutant, we transformed the ∆noxA strain with the deletion construct of yap1. Validation of all knockout strains was achieved with PCR, using genespecific primers (Table S3,  The coding sequence of the yap1 gene, flanked by~2.0 kb-long genomic regions, was amplified from genomic DNA of V. dahliae 123V and co-transformed into protoplasts of the ∆yap1 strain with plasmid pUCATPH [50] (which carries the hph cassette conferring resistance to hygromycin B) to generate the complemented strain yap1-c.

Characterization of Morphology, Physiology, and Stress Response of V. dahliae Strains
Morphological and physiological characterization of fungal strains, as well as assessment of their stress sensitivity, were performed as previously described [47], with a minor modification for the determination of germination frequency. In particular, strains were grown for 12 h in CM and then checked microscopically for the emergence of germ tubes.
To characterize stress tolerance and responses, we exposed fungal strains to a variety of oxidative agents (H 2 O 2 , paraquat, iprodione, and N-acetyl cysteine), substances that induce osmotic stress (NaCl and sorbitol), cell wall damaging factors (amphotericin B obtained from Biosera, Nuaille, France, fluconazole from Pfizer, Brooklyn, NY, USA, calcofluor white M2R, and Congo red), trace elements (CaCl 2 , CuSO 4 , and FeSO 4 ), and an inhibitor of TOR kinase (sirolimus from Cayman Chemical, Ann Arbor, MI, USA). In addition, we used fungicides that target complex I (sodium amytal), complex II (isopyrazam, flutolanil), complex III (azoxystrobin, kresoxim-methyl, pyraclostrobin-QoI/cytB site; annisulborn-Qi site), and complex IV (sodium cyanide) of the electron transport chain to investigate possible interactions between the ROS-generating systems. All chemicals were purchased from Sigma-Aldrich, St. Louis, MO, USA, unless otherwise specified above. According to our previously described methods [47], we used spot dilution assays and determined relative growth inhibition (i.e., % growth inhibition = ((colony diameter on CM − colony diameter in stress condition)/(colony diameter on CM)) × 100). All experiments were performed at least in triplicate.

Virulence Assays
Plant pathogenicity bioassays were performed according to our previously published methods [47,51]. Briefly, eggplant seedlings were inoculated with the corresponding fungal strain by soil drenching (20 mL of conidial suspension of 5.0 × 10 6 conidia/mL per pot). Plants were maintained at 24 • C with a 12 h light/dark cycle. Assessment of disease severity at each time point (up to 40 days) and determination of relative AUDPC and plant growth parameters were performed according to previously described protocols [51,52]. We re-isolated fungal strains from the treated plants by transferring xylem chips (three chips from nine randomly selected plants per treatment) to acidified PDA plates. The isolation ratio was expressed as the number of xylem chips from each treatment that exhibited fungal growth.

Cellophane Penetration Assays
The ability of V. dahliae strains to penetrate cellophane membranes was assessed as follows. Conidial suspensions from the examined strains were collected from 7-day-old PDA cultures, and 5 × 10 6 conidia of each were transferred onto CM plates overlaid with sterile cellophane sheets. The samples were incubated at 24 • C for five days. Cellophane sheets were then removed, and the plates were incubated for four additional days before being scored for fungal growth. For the microscopic examination of hyphopodial formation, the same membranes (i.e., after five days of incubation with the respective conidial suspension) were washed with sterile water and observed under a microscope.

Superoxide anion radicals (O 2
•− ) in germlings and in the mycelium of V. dahliae strains were detected by a nitro blue tetrazolium chloride (NBT; Cayman Chemical, Ann Arbor, MI, USA) staining assay. Conidia of each strain were grown for 16 h in CM, in plate wells with sterile coverslips at their bottom. Following incubation, the medium was removed, 5 mL of a 0.2% NBT solution was added, and the samples were incubated at room temperature (RT), in the dark, for 45 min. The coverslips were then washed with ethanol and checked microscopically. Detection of superoxide anion radicals in the mycelium was performed with the addition of 10 mL of a 0.2% NBT solution on 20-day-old cultures grown on CM. Plates were incubated for 45 min at RT, which was followed by washing with ethanol and incubation for another 30 min, before being air-dried and scored. Staining assays were performed in duplicate.

Reverse Transcription Quantitative PCR (RT-qPCR)
Determination of the expression levels of selected genes was performed as follows. Wild-type, ∆noxA, ∆yap1, and ∆noxA ∆yap1 strains were grown for five days in CM at 24 • C. Mycelia were collected, washed with water, and when desired treated with 1.5 mM H 2 O 2 for 45 min. They were then snap-frozen in liquid nitrogen (N 2 ), ground to a fine powder under N 2 , and used for extraction of total RNA with the NucleoSpin RNA kit (Macherey-Nagel, Düren, Germany). Total RNA was reverse transcribed to cDNA using the PrimeScript cDNA synthesis kit (Takara Bio, Kusatsu, Japan). Quantitative PCR (qPCR) was performed using the KAPA SYBR Fast Universal Master Mix (Roche, Basel, Switzerland) in a Mx3000P real-time PCR instrument (Stratagene California, San Diego, CA, USA), according to the manufacturer's instructions. For each primer pair (Table S3, Supplementary Materials), we performed optimization of oligo concentrations and construction of standard curves in preliminary experiments. All curves were highly linear (R 2 > 0.999), and their amplification efficiencies ranged between 95% and 105%. The cycling protocol consisted of an initial denaturation step at 95 • C for 5 min, followed by 40 cycles of 10 s at 95 • C (denaturation), 20 s at 60 • C (annealing), and 20 s at 72 • C (elongation). Dissociation curves were generated for each primer pair with the following protocol. Samples were first incubated for 1 min at 95 • C, then at 55 • C for 30 s, and finally a temperature ramp (0.1 • C/s) up to 95 • C was applied with continuous collection of fluorescence readings. A single product was amplified with each primer pair. No-template control samples were included in each run to check for contamination and significant formation of primer dimers.
The 2 −∆∆Ct method [53] was used for analysis of the results, with the modification that each ∆Ct value was calculated as the difference between the Ct values of the reference and target genes (i.e., Ct reference-Ct target), and the fold change was calculated by the formula 2 ∆∆Ct . The V. dahliae β-tubulin gene (VDAG_10074) was used as the internal reference in all experiments. The ∆∆Ct value and fold change for each gene in different mutants and/or conditions was calculated relative to the control condition (i.e., expression level of the corresponding gene in the untreated wild-type strain). Student's t-tests were used to assess differences of ∆∆Ct values. Three biological (i.e., independent cultures) and two technical replicates were performed for each strain and condition.

Quantification of CAT-Mediated Fusion
We used our previously optimized and described methods for reproducible quantification of CAT-mediated self and nonself fusion [23,54] to investigate the possible involvement of noxA and yap1 in somatic cell fusion. Briefly, conidia from 7-day-old PDA cultures were collected and transferred to 6-mm Petri dishes, with coverslips at their bottom (containing 5 mL of CAT medium each) to a final concentration of 2.0 × 10 6 conidia/mL. Plates were incubated for 60 h at 24 • C before imaging. Each strain/pairing was tested in triplicate and 200 fusions were recorded per replicate.

Microscopy
Microscopic examination of fungal germlings, hyphae, and NBT-stained samples, as well as investigation of CAT-mediated fusion, were performed using a Zeiss Axioplan epifluorescence microscope equipped with a differential interference contrast (DIC) optical system, a set of filters BP450-490 (excitation) and BP515-595 (emission), and a Zeiss Axiocam MRc5 digital camera. Methylene blue staining (0.005% w/v) was used to differentiate live from dead cells; samples were incubated at 25 • C for 5 min (in the dark) before imaging.

Roles of NoxA and Yap1 in V. dahliae Morphogenesis and Physiology
The NADPH oxidase A (NoxA) and the transcriptional regulator Yap1 are key components of ROS metabolism in fungi and other organisms. The single homolog of noxA in the V. dahliae genome (VDAG_06812 in the reference genome of strain Ls.17) codes for a predicted protein of 555 aa with a high similarity to its N. crassa Nox1 homolog (99% query coverage, 90% sequence similarity), and we have recently described its deletion in the wild-type V. dahliae strain 123V [23]. The A. nidulans Yap1 protein sequence was used as a query in tBlastN genomic searches for the identification of its single V. dahliae homolog, which encodes a predicted protein of 583 aa (VDAG_01588; protein similarity: 52%). This protein is predicted to contain the characteristic bzip domain and the signal peptides of its S. cerevisiae homolog, as well as the conserved cysteine residues that are essential for its nuclear localization ( Figure S1, Supplementary Materials). For the functional analysis of yap1 in V. dahliae, the gene was deleted from strain 123V, and the resulting deletion mutant (∆yap1) was validated by PCR and Southern blot analyses ( Figure S2, Supplementary Materials). The wild-type yap1 gene was re-introduced into ∆yap1 to generate the complemented strain yap1-c. A double deletion ∆noxA ∆yap1 mutant was constructed via double homologous recombination following transformation of ∆noxA conidia with the yap1 deletion construct ( Figure S2, Supplementary Materials).
Morphogenesis and physiology of strains ∆noxA, ∆yap1, and ∆noxA ∆yap1 were compared to those of their wild type on growth media PDA, CM, and MM. Production of microsclerotia was significantly reduced in mutants ∆noxA and ∆noxA ∆yap1 on PDA, while it was almost absent from ∆yap1 on all media ( Figure 1A). The ability of ∆noxA and ∆noxA ∆yap1 to produce aerial hyphae was compromised on CM and MM ( Figure 1A), and these strains also exhibited moderately slower growth on PDA (p < 0.01 and p < 0.05, respectively; Figure 1B). In contrast, deletion of yap1 did not affect the formation of aerial mycelium, but it slightly reduced its growth rate (p < 0.05, Figure 1A,B).
Regarding development of conidia, a non-significant increase in their production was observed in ∆noxA, but their ability to germinate was drastically limited in this mutant (51.3% reduction), similarly to what was observed in ∆noxA ∆yap1 (43.7% reduction) ( Figure 1B). On the other hand, ∆yap1 exhibited a 10-fold reduction in conidiation ( Figure 1B), although its conidia germinated normally ( Figure 1B,C). We also observed that mature hyphae of ∆noxA and ∆noxA ∆yap1 were significantly thinner than those of the wild type ( Figure 1D,E). All these defects were fully rescued in the corresponding complemented strains ( Figure 1A,B,E). resulting deletion mutant (Δyap1) was validated by PCR and Southern blot analyses ( Figure S2, Supplementary Materials). The wild-type yap1 gene was re-introduced into Δyap1 to generate the complemented strain yap1-c. A double deletion ΔnoxA Δyap1 mutant was constructed via double homologous recombination following transformation of ΔnoxA conidia with the yap1 deletion construct ( Figure S2, Supplementary Materials). Morphogenesis and physiology of strains ΔnoxA, Δyap1, and ΔnoxA Δyap1 were compared to those of their wild type on growth media PDA, CM, and MM. Production of microsclerotia was significantly reduced in mutants ΔnoxA and ΔnoxA Δyap1 on PDA, while it was almost absent from Δyap1 on all media ( Figure 1A). The ability of ΔnoxA and ΔnoxA Δyap1 to produce aerial hyphae was compromised on CM and MM ( Figure 1A), and these strains also exhibited moderately slower growth on PDA (p < 0.01 and p < 0.05, respectively; Figure 1B). In contrast, deletion of yap1 did not affect the formation of aerial mycelium, but it slightly reduced its growth rate (p < 0.05, Figure 1A,B).
Regarding development of conidia, a non-significant increase in their production was observed in ΔnoxA, but their ability to germinate was drastically limited in this mutant (51.3% reduction), similarly to what was observed in ΔnoxA Δyap1 (43.7% reduction) ( Figure 1B). On the other hand, Δyap1 exhibited a 10-fold reduction in conidiation ( Figure 1B), although its conidia germinated normally ( Figure 1B,C). We also observed that mature hyphae of ΔnoxA and ΔnoxA Δyap1 were significantly thinner than those of the wild type ( Figure 1D,E). All these defects were fully rescued in the corresponding complemented strains ( Figure 1A,B,E).  For each strain, the widths of at least 50 non-apical hyphal compartments from colony areas of the same age were recorded. Bars: SD. Statistical significance of differences from the wild type was assessed using Student's t-tests (*** p ≤ 0.001).

NoxA, but Not Yap1, Is Essential for the Penetration Ability and Pathogenicity of V. dahliae
To investigate the possible involvement of NoxA and Yap1 in V. dahliae pathogenicity, we assessed the ability of the deletion mutants to cause disease in eggplant, in comparison to their wild type. Both ∆noxA and ∆noxA ∆yap1 mutants exhibited diminished pathogenicity, with 88% and 76% of the treated plants (respectively) remaining completely asymptomatic 40 days after inoculation, while the remaining plants showed only weak symptoms (Figure 2A,B). In contrast, deletion of yap1 only slightly limited the fungal potential to cause disease (Figure 2A,B). Complementation of the corresponding mutants with the wild-type alleles of noxA and yap1 fully rescued the ability of the fungus to cause severe disease, characterized by defoliation, wilting, and chlorosis ( Figure 2A,B).
corresponding mutants with the wild-type alleles of noxA and yap1 fully rescued the ability of the fungus to cause severe disease, characterized by defoliation, wilting, and chlorosis ( Figure 2A,B).
We further performed a time-course analysis of virulence to gain a better understanding of the involvement of noxA and yap1 in the induction of plant disease ( Figure 2C). The wild-type strain started causing observable symptoms in the infected plants 12 days after inoculation, reaching a mean disease severity of 55.5% (±7.3%) at the end of the experiment (29 days), and an overall relative AUDPC value of 27.5% (±3.9%) ( Figure 2C,D). Deletion of noxA completely prevented the expression of disease symptoms during this time period, similarly to the ΔnoxA Δyap1 strain, which achieved a relative AUDPC score of 0.1% (±0.1%) ( Figure 2C,D). In contrast, only a minor reduction was detected in virulence of Δyap1 ( Figure 2C,D). These findings were further supported by the determination of the average plant fresh weight at the end of the experiment, with ΔnoxA and ΔnoxA Δyap1 having no effect, whereas Δyap1 led to a reduction similar to that of the wild-type strain ( Figure 2E). Consistently, our attempts to re-isolate the fungus from xylem chips (29 days post-inoculation) revealed a significantly reduced presence of the pathogen in the xylem vessels of plants inoculated with ΔnoxA or ΔnoxA Δyap1, in contrast to Δyap1, which achieved systemic colonization at wild-type levels ( Figure 2F). We further performed a time-course analysis of virulence to gain a better understanding of the involvement of noxA and yap1 in the induction of plant disease ( Figure 2C). The wild-type strain started causing observable symptoms in the infected plants 12 days after inoculation, reaching a mean disease severity of 55.5% (±7.3%) at the end of the experiment (29 days), and an overall relative AUDPC value of 27.5% (±3.9%) ( Figure 2C,D). Deletion of noxA completely prevented the expression of disease symptoms during this time pe-riod, similarly to the ∆noxA ∆yap1 strain, which achieved a relative AUDPC score of 0.1% (±0.1%) ( Figure 2C,D). In contrast, only a minor reduction was detected in virulence of ∆yap1 ( Figure 2C,D). These findings were further supported by the determination of the average plant fresh weight at the end of the experiment, with ∆noxA and ∆noxA ∆yap1 having no effect, whereas ∆yap1 led to a reduction similar to that of the wild-type strain ( Figure 2E). Consistently, our attempts to re-isolate the fungus from xylem chips (29 days post-inoculation) revealed a significantly reduced presence of the pathogen in the xylem vessels of plants inoculated with ∆noxA or ∆noxA ∆yap1, in contrast to ∆yap1, which achieved systemic colonization at wild-type levels ( Figure 2F).
The inability of the ∆noxA and ∆noxA ∆yap1 strains to cause disease in their plant hosts, as well as their severely reduced presence in the xylem of inoculated plants, led us to the hypothesis that noxA is possibly necessary for efficient root penetration. We therefore investigated the capacity of the deletion mutants to penetrate cellophane membranes by inoculating cellophane sheets overlaying CM plates with conidia of each strain and incubating them for 5 days. Cellophane sheets were then removed, plates were incubated for another 4 days, and they were finally scored for fungal growth. In support of our hypothesis, deletion of noxA, or both noxA and yap1, completely abolished penetration of cellophane, in contrast to the wild-type, the ∆yap1, and the complemented strains, which grew profoundly on but also below the membrane sheets ( Figure 3A). Furthermore, microscopic examination of the removed membranes revealed a drastic reduction in the frequency of normally developed hyphopodia (i.e., infectious structures involved in host penetration [43]) in ∆noxA and ∆noxA ∆yap1, whereas all other strains retained the ability to form numerous hyphopodia ( Figure 3B). Our results indicate that NoxA, but not Yap1, plays a major role in plant penetration and has essential functions in V. dahliae pathogenicity. The inability of the ΔnoxA and ΔnoxA Δyap1 strains to cause disease in their plant hosts, as well as their severely reduced presence in the xylem of inoculated plants, led us to the hypothesis that noxA is possibly necessary for efficient root penetration. We therefore investigated the capacity of the deletion mutants to penetrate cellophane membranes by inoculating cellophane sheets overlaying CM plates with conidia of each strain and incubating them for 5 days. Cellophane sheets were then removed, plates were incubated for another 4 days, and they were finally scored for fungal growth. In support of our hypothesis, deletion of noxA, or both noxA and yap1, completely abolished penetration of cellophane, in contrast to the wild-type, the Δyap1, and the complemented strains, which grew profoundly on but also below the membrane sheets ( Figure 3A). Furthermore, microscopic examination of the removed membranes revealed a drastic reduction in the frequency of normally developed hyphopodia (i.e., infectious structures involved in host penetration [43]) in ΔnoxA and ΔnoxA Δyap1, whereas all other strains retained the ability to form numerous hyphopodia ( Figure 3B). Our results indicate that NoxA, but not Yap1, plays a major role in plant penetration and has essential functions in V. dahliae pathogenicity.

NoxA Is Involved in ROS Metabolism and Its Oxidative Activity Is Counteracted by Yap1
Staining of O2 •− radicals using NBT revealed their accumulation mostly in the apical parts of germlings and at the periphery of colonies of the ΔnoxA mutant, in contrast to their more uniform presence in the wild-type strain ( Figure 4A). Peripheral staining of colonies was also observed in ΔnoxA Δyap1, whereas the wild-type pattern was restored in the complemented strain ( Figure 4A). These observations support the hypothesis that NoxA is involved in ROS metabolism in V. dahliae. On the other hand, deletion of yap1

NoxA Is Involved in ROS Metabolism and Its Oxidative Activity Is Counteracted by Yap1
Staining of O 2 •− radicals using NBT revealed their accumulation mostly in the apical parts of germlings and at the periphery of colonies of the ∆noxA mutant, in contrast to their more uniform presence in the wild-type strain ( Figure 4A). Peripheral staining of colonies was also observed in ∆noxA ∆yap1, whereas the wild-type pattern was restored in the complemented strain ( Figure 4A). These observations support the hypothesis that NoxA is involved in ROS metabolism in V. dahliae. On the other hand, deletion of yap1 resulted in generally increased formazan precipitation ( Figure 4A), which indicates de-fects in ROS metabolism and is consistent with the conserved role of this regulator in ROS detoxification.
We further characterized the roles of NoxA and Yap1 in response to oxidative stress by exposing the wild-type and knockout strains to the oxidizing agent H 2 O 2 , the herbicide paraquat, the fungicide iprodione, and the ROS scavenger N-acetyl cysteine (NAC). As expected, we detected significant growth defects of ∆yap1 in the presence of any ROSinducing agent ( Figure 4B,C). On the other hand, ∆noxA was inhibited to a lesser extent by H 2 O 2 (at high concentrations), iprodione, and the ROS-scavenger NAC ( Figure 4B,C). Notably, the double deletion mutant ∆noxA ∆yap1 exhibited the same behavior as ∆noxA ( Figure 4B,C), indicating an epistatic relationship between the two genes. resulted in generally increased formazan precipitation ( Figure 4A), which indicates defects in ROS metabolism and is consistent with the conserved role of this regulator in ROS detoxification. We further characterized the roles of NoxA and Yap1 in response to oxidative stress by exposing the wild-type and knockout strains to the oxidizing agent H2O2, the herbicide paraquat, the fungicide iprodione, and the ROS scavenger N-acetyl cysteine (NAC). As expected, we detected significant growth defects of Δyap1 in the presence of any ROSinducing agent ( Figure 4B,C). On the other hand, ΔnoxA was inhibited to a lesser extent by H2O2 (at high concentrations), iprodione, and the ROS-scavenger NAC ( Figure 4B,C). Notably, the double deletion mutant ΔnoxA Δyap1 exhibited the same behavior as ΔnoxA ( Figure 4B,C), indicating an epistatic relationship between the two genes. To gain a better understanding of the roles of V. dahliae noxA and yap1 in the OSR, we investigated their expression patterns before and after treatment with H2O2 by RT-qPCR ( Figure 5). The noxA gene is strongly induced upon oxidative stress in the wild type, and neither its induction pattern nor its transcript levels were affected by deletion of yap1. On the other hand, yap1, which is down-regulated under oxidative stress in the wild type, failed to reach its normal expression levels in the ΔnoxA mutant background under these conditions. When we also analyzed the expression levels of the stress-activated Hog1 MAP kinase gene before treatment with H2O2, we observed significantly lower levels in all knockout mutants than in the wild type. Under oxidative stress, however, hog1 was down-regulated in the wild type, followed a similar but less pronounced trend in Δyap1, and moderate induction in ΔnoxA and ΔnoxA Δyap1 ( Figure 5).
Furthermore, we characterized the responses to oxidative stress of three important downstream genes known to be involved in the OSR, i.e., catalase (cat1), superoxide dismutase (sod1), and glutathione reductase (glr1). As expected, induction of all three was To gain a better understanding of the roles of V. dahliae noxA and yap1 in the OSR, we investigated their expression patterns before and after treatment with H 2 O 2 by RT-qPCR ( Figure 5). The noxA gene is strongly induced upon oxidative stress in the wild type, and neither its induction pattern nor its transcript levels were affected by deletion of yap1. On the other hand, yap1, which is down-regulated under oxidative stress in the wild type, failed to reach its normal expression levels in the ∆noxA mutant background under these conditions. When we also analyzed the expression levels of the stress-activated Hog1 MAP kinase gene before treatment with H 2 O 2 , we observed significantly lower levels in all knockout mutants than in the wild type. Under oxidative stress, however, hog1 was down-regulated in the wild type, followed a similar but less pronounced trend in ∆yap1, and moderate induction in ∆noxA and ∆noxA ∆yap1 ( Figure 5).
Furthermore, we characterized the responses to oxidative stress of three important downstream genes known to be involved in the OSR, i.e., catalase (cat1), superoxide dismutase (sod1), and glutathione reductase (glr1). As expected, induction of all three was detected in the wild-type strain in the presence of H 2 O 2 , whereas they were expressed at lower levels and failed to be induced by oxidative stress in the absence of their transcriptional regulator Yap1 ( Figure 5). Interestingly, we detected significantly higher mRNA levels of cat1 and glr1 in the ∆noxA mutant than in the wild type under normal conditions, while sod1 and glr1 failed to reach their wild-type levels under oxidative stress in ∆noxA ( Figure 5). The ∆noxA ∆yap1 strain exhibited a similar expression pattern to ∆yap1 for cat1 and glr1, while the expression of sod1 was up-regulated under normal conditions in this strain ( Figure 5). Our findings indicate that V. dahliae NoxA is up-regulated by oxidative stress and implicate it in the transcriptional regulation of both the OSR and the important stress-responsive MAP kinase Hog1. detected in the wild-type strain in the presence of H2O2, whereas they were expressed at lower levels and failed to be induced by oxidative stress in the absence of their transcriptional regulator Yap1 ( Figure 5). Interestingly, we detected significantly higher mRNA levels of cat1 and glr1 in the ΔnoxA mutant than in the wild type under normal conditions, while sod1 and glr1 failed to reach their wild-type levels under oxidative stress in ΔnoxA ( Figure 5). The ΔnoxA Δyap1 strain exhibited a similar expression pattern to Δyap1 for cat1 and glr1, while the expression of sod1 was up-regulated under normal conditions in this strain ( Figure 5). Our findings indicate that V. dahliae NoxA is upregulated by oxidative stress and implicate it in the transcriptional regulation of both the OSR and the important stress-responsive MAP kinase Hog1. Three biological replicates (and two technical replicates for each) were performed and analyzed for each strain and gene. In the ΔΔCt plots, the horizontal bars represent mean ΔΔCt values and the bars represent the SD between the three biological replicates. Statistical significance of differences from the untreated wild-type strain was tested by Student's t-tests (* p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001).

NoxA Is Involved in the Response to Cell Wall Stress and Possibly Interacts with Components of Important Biological Processes
We further investigated the possible involvement of NoxA and Yap1 in fungal responses to other sources of cellular stress by characterizing the behavior of the corresponding mutant strains in the presence of substances inducing hyperosmotic stress, cell wall-perturbating agents, high concentrations of trace metals, and various fungicides ( Figure 6A,B). Regarding osmotic stress, only a minor growth inhibition of ΔnoxA and Δyap1 was recorded in the presence of NaCl. In contrast, cell wall stress significantly compromised the growth rates (calcofluor white and Congo red) and the germination frequency (calcofluor) of ΔnoxA and ΔnoxA Δyap1, whereas Δyap1 exhibited limited sensitivity only to calcofluor. Consistently, the drug amphotericin B, which disorganizes fungal cell membranes by targeting ergosterol, and also causes oxidative stress, impaired growth and germination of ΔnoxA and ΔnoxA Δyap1. However, the inhibitor of ergosterol biosynthesis fluconazole specifically restricted growth of Δyap1 and, to a lesser extent, that Three biological replicates (and two technical replicates for each) were performed and analyzed for each strain and gene. In the ∆∆Ct plots, the horizontal bars represent mean ∆∆Ct values and the bars represent the SD between the three biological replicates. Statistical significance of differences from the untreated wild-type strain was tested by Student's t-tests (* p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001).

NoxA Is Involved in the Response to Cell Wall Stress and Possibly Interacts with Components of Important Biological Processes
We further investigated the possible involvement of NoxA and Yap1 in fungal responses to other sources of cellular stress by characterizing the behavior of the corresponding mutant strains in the presence of substances inducing hyperosmotic stress, cell wall-perturbating agents, high concentrations of trace metals, and various fungicides ( Figure 6A,B). Regarding osmotic stress, only a minor growth inhibition of ∆noxA and ∆yap1 was recorded in the presence of NaCl. In contrast, cell wall stress significantly compromised the growth rates (calcofluor white and Congo red) and the germination frequency (calcofluor) of ∆noxA and ∆noxA ∆yap1, whereas ∆yap1 exhibited limited sensitivity only to calcofluor. Consistently, the drug amphotericin B, which disorganizes fungal cell membranes by targeting ergosterol, and also causes oxidative stress, impaired growth and germination of ∆noxA and ∆noxA ∆yap1. However, the inhibitor of ergosterol biosynthesis fluconazole specifically restricted growth of ∆yap1 and, to a lesser extent, that of the double deletion mutant. Motivated by the observation that deletion of yap1 in yeast causes decreased resistance to metals [55], we checked resistance of the V. dahliae noxA and yap1 mutants to Ca 2+ , Cu 2+ , and Fe 2+ cations ( Figure S3A,B, Supplementary Materials). While Ca 2+ and Cu 2+ indeed restricted radial growth of ∆yap1, Fe 2+ had a more pronounced effect on the noxA-deficient strains.
Apart from the Nox complexes, mitochondria are a major intracellular source of ROS. We tested fungicides that target the four complexes (I-IV) of the respiratory chain (i.e., azoxystrobin, sodium cyanide, pyraclostrobin, kresoxim-methyl, annisulborn, sodium amytal, flutolanil, and isopyrazam), and we found them all to cause significant growth reduction in ∆noxA and ∆noxA ∆yap1 strains, while ∆yap1 was not affected (Figure 6A,B; Figure S3A,B, Supplementary Materials). These findings are suggestive of a possible crosstalk between the two major cellular sources of ROS.
Finally, since NoxA has been shown to interact with the machinery of autophagy in phagocytic cells [56], we tested the responses of the V. dahliae mutants to rapamycin, an inhibitor of the TORC1 kinase and, therefore, an inducer of autophagy ( Figure 6A,B). Deletion of noxA, but not of yap1, led to significantly increased sensitivity to the inhibitor, which implies a possible dysregulation of autophagy.  Figure S3A,B, Supplementary Materials). While Ca 2+ and Cu 2+ indeed restricted radial growth of Δyap1, Fe 2+ had a more pronounced effect on the noxA-deficient strains. Apart from the Nox complexes, mitochondria are a major intracellular source of ROS. We tested fungicides that target the four complexes (I-IV) of the respiratory chain (i.e., azoxystrobin, sodium cyanide, pyraclostrobin, kresoxim-methyl, annisulborn, sodium amytal, flutolanil, and isopyrazam), and we found them all to cause significant growth reduction in ΔnoxA and ΔnoxA Δyap1 strains, while Δyap1 was not affected (Figure 6A,B; Figure S3A,B, Supplementary Materials). These findings are suggestive of a possible crosstalk between the two major cellular sources of ROS.
Finally, since NoxA has been shown to interact with the machinery of autophagy in phagocytic cells [56], we tested the responses of the V. dahliae mutants to rapamycin, an inhibitor of the TORC1 kinase and, therefore, an inducer of autophagy ( Figure 6A,B). Deletion of noxA, but not of yap1, led to significantly increased sensitivity to the inhibitor, which implies a possible dysregulation of autophagy. Figure 6. Effects of osmotic stress, cell wall-and plasma membrane-perturbating agents, and antifungal agents that target the respiratory chain on V. dahliae ΔnoxA and Δyap1 mutants: (A) effects of NaCl, sorbitol, calcofluor white M2R, Congo red, amphotericin B, fluconazole, azoxystrobin, sodium cyanide, pyraclostrobin, kresoxim-methyl, and rapamycin on conidial germination (growth for 3 days). (B) Relative growth inhibition of colonies caused by the same substances. Bars: SD. Statistical significance of differences from the wild type was tested by Student's t-tests (* p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001). All concentrations are expressed in μg/mL, except for NaCl (M), sorbitol (M), and rapamycin (nM). All experiments were performed in triplicate. Figure 6. Effects of osmotic stress, cell wall-and plasma membrane-perturbating agents, and antifungal agents that target the respiratory chain on V. dahliae ∆noxA and ∆yap1 mutants: (A) effects of NaCl, sorbitol, calcofluor white M2R, Congo red, amphotericin B, fluconazole, azoxystrobin, sodium cyanide, pyraclostrobin, kresoxim-methyl, and rapamycin on conidial germination (growth for 3 days). (B) Relative growth inhibition of colonies caused by the same substances. Bars: SD. Statistical significance of differences from the wild type was tested by Student's t-tests (* p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001). All concentrations are expressed in µg/mL, except for NaCl (M), sorbitol (M), and rapamycin (nM). All experiments were performed in triplicate.

NoxA, but Not Yap1, Is Essential in Both Partners for CAT-Mediated Cell Fusion
We have previously found that NoxA in V. dahliae is essential for somatic fusion of conidia or germlings via CATs [23]. On the contrary, deletion of yap1 exerted no effect on CAT-mediated fusion ( Figure 7A). The double deletion mutant exhibited the same behavior as ∆noxA ( Figure 7A). Surprisingly, the ∆noxA mutant retained its capacity for hyphal fusion between mature hyphae ( Figure 7B), which suggests differences between the regulatory mechanisms of conidial and hyphal fusion.

Discussion
The dual nature of ROS in biological systems as a potential cause of cellular damage on the one hand, and as signaling components of important developmental processes on the other, has been the focus of extensive research [1,2,5,8]. This duality is clearly illustrated in the case of fungal plant pathogens, which need to cope with their hostderived ROS bursts during infection [4], while they generate and use endogenous ROS for their own development and pathogenicity [5,6]. Therefore, the survival and ecological success of these organisms depend on their ability to maintain a redox balance that presumably requires coordination of ROS generation and scavenging systems. In this study, we investigated the roles of the ROS-producing NADPH oxidase NoxA in the biology of the important plant-pathogenic fungus V. dahliae, in comparison to its We hypothesized that the role of NoxA in cell fusion via CATs could be related to the generation or perception of the unknown signal that is involved in pre-fusion cell communication. If that process were unidirectional (i.e., each fusion partner emitting or receiving the signal, but not both), a functional copy of noxA in only one of the interacting partners could be sufficient for fusion of a subset of cells. To test this possibility, we paired the deletion strains with the strain Ls.17 H1-sgfp (expressing GFP-tagged histone H1 in its nuclei), which permitted the microscopic identification of nonself fusions. The presence of functional noxA genes in both partners was found to be essential for their fusion ( Figure 7C); deletion of yap1 from one partner showed no effect, as expected.
It is unclear whether NoxA functions in cell fusion by causing localized ROS production with direct effects on partner recognition or the commitment to fuse, or indirectly by transducing signals to alter the expression of other genes or the activity of their products. We found that, during the homing phase of the interacting cells prior to their fusion, O 2 •− always accumulated at the tips of the mutually attracted CATs (Figure 7D), suggesting that localized generation of ROS putatively by NoxA is indeed involved in the process. No ROS accumulation was detected after completion of CAT-mediated fusion ( Figure 7D). On the other hand, deletion of noxA had no effect on the expression levels of the MAP kinases Fus3 and Slt2, which are important components of the pre-fusion communication mechanism in N. crassa [22], and are also required for CAT fusion in V. dahliae [23].
Fusion of vegetatively incompatible strains of fungi, including V. dahliae, usually triggers an incompatibility reaction that can cause cell death of the anastomosed compartments [54]. Based on the observations that this reaction in N. crassa is characterized by an induction of ROS [57], and that genes involved in the OSR are up-regulated during CAT formation in Colletotrichum gloeosporioides [58], we used NBT staining to study O 2 •− accumulation in incompatible V. dahliae fusions (pairing Ls.17 H1-sgfp × 123V). One-third of such fusions (n = 30) exhibited increased formazan precipitation ( Figure 7F), indicating that ROS accumulation is indeed associated with the onset of the incompatibility reaction, similarly to what was observed in N. crassa [57]. Deletion of yap1 had no effect on the frequency of inviable fusions (Ls.17 H1-sgfp × 123V ∆yap1 pairing), which suggests either that the observed ROS accumulation does not mediate the catastrophic reaction or that Yap1 does not respond to it.

Discussion
The dual nature of ROS in biological systems as a potential cause of cellular damage on the one hand, and as signaling components of important developmental processes on the other, has been the focus of extensive research [1,2,5,8]. This duality is clearly illustrated in the case of fungal plant pathogens, which need to cope with their host-derived ROS bursts during infection [4], while they generate and use endogenous ROS for their own development and pathogenicity [5,6]. Therefore, the survival and ecological success of these organisms depend on their ability to maintain a redox balance that presumably requires coordination of ROS generation and scavenging systems. In this study, we investigated the roles of the ROS-producing NADPH oxidase NoxA in the biology of the important plantpathogenic fungus V. dahliae, in comparison to its transcription factor Yap1, a conserved central regulator of ROS detoxification [13,28]. We found that NoxA has multifaceted roles in important developmental and physiological processes, plays an essential role in virulence, mediates responses to different types of cellular stress, and is required for somatic cell fusion.
Fungal NADPH oxidase complexes have been attributed pleiotropic functions in morphogenesis and pathogenicity [9,12,13,59]. In V. dahliae, we found NoxA to be required for normal formation of microsclerotia, i.e., dormant resting structures that are crucial for long-term survival and spread of the species [42]. In addition, V. dahliae NoxA is necessary for normal formation of aerial hyphae and germination of its asexual spores, in agreement with similar previously described roles in, e.g., N. crassa, Epichloë festucae, C. purpurea, and S. macrospora [16][17][18]60]. Furthermore, our investigation revealed that NoxA is essential for virulence in V. dahliae, and our findings suggest that this can be at least partly due to its involvement in the formation of hyphopodia, i.e., infectious swollen hyphae that can penetrate the roots of the plant hosts to initiate the disease cycle [43,44]. Although the regulatory mechanisms that underlie the development of V. dahliae hyphopodia are mostly unknown, recent reports implicated in this process both Ste11, a component of MAP kinase signaling, and Csin1, a cellophane surface-induced protein that acts through the cAMP pathway [61,62]. These findings suggest putative interactions of ROS with important signaling transduction pathways. Previous studies have demonstrated the involvement of NoxA in the formation of appressoria in the plant pathogen M. oryzae [10,11] and traps in the nematode pathogen Arthrobotrys oligospora [63]. These findings together indicate conserved involvement of the ROS-generating enzyme NoxA in the early stages of infection of diverse pathogenic fungi.
Another explanation for the highly compromised ability of V. dahliae to cause systemic infection and disease symptoms in the absence of NoxA could be its increased sensitivity to various types of environmental stress, most notably cell wall and oxidative stress. Oxidative bursts of ROS have been described to play a role in cell wall biosynthesis [9], which could be linked to the sensitivity of the V. dahliae ∆noxA mutant to cell wall-perturbating agents. Our analyses also revealed reduced resistance of ∆noxA to oxidative stress, which could be attributed to dysregulation of antioxidant genes, such as the example of sod1 that we observed in our study. In addition, deletion of noxA resulted in dysregulation of the important stress-activated MAP kinase Hog1, which could be directly or indirectly linked to some of the identified pleiotropic phenotypes of NoxA. These observations could be likely attributed to the expected function of ROS as secondary signals that could potentially control the expression and function of multiple transcription factors and signaling components via post-translational modification [2]. A possible mechanism for this could involve the control of cellular localization of key signaling components, such as the MAP kinase Mpk1 in P. anserina, whose normal localization depends on nox1 [64]. Notably, the absence of functional NoxA rendered the fungus sensitive to several antifungal agents, including representatives that block the central TOR kinase, as well as the four complexes of the mitochondrial electron transport chain. The latter observation indicates that the maintenance of a minimum level of intracellular ROS is important for cell functionality and viability. It could also be relevant to the crosstalk that has been demonstrated between mitochondrial ROS and NADPH oxidases in mammalian cells [65], hypothetically reflecting the function of mechanisms that ensure the coordination of those two major intracellular sources of ROS in the context of oxidative homeostasis.
Conidial and hyphal fusion is an integral component of the establishment and development of fungal colonies [66], and nonself fusion could possibly grant access to parasexual generation of diversity [23,54]. We have previously demonstrated that NoxA is essential for CAT-mediated fusion of V. dahliae conidia or germlings [23], and we further found in the present study that the requirement for functional NoxA applies to both interacting cells. We also detected O 2 •− accumulation at the tips of wild-type germlings during their homing phase of interaction prior to fusion, which could most likely be attributed to the activity of NoxA. Similar observations have previously been made in B. cinerea [25]. We hypothesize that temporally and spatially regulated ROS bursts could be involved in the pre-fusion communication of the interacting cells, possibly linked to the oscillatory recruitment of signaling components (such as the MAP kinase Fus3 and the scaffold protein SOFT) to cell tips that controls cell fusion in N. crassa [67].
The transcription factor Yap1 has a conserved role in ROS detoxification among diverse organisms as a central regulatory element that participates in the orchestration of their OSR [30][31][32][33][34][35][36][37]. Indeed, we found deletion of yap1 to result in increased sensitivity to all oxidative agents tested, as well as to the fungicide fluconazole, which can induce ROS production in Cryptococcus neoformans [68,69]. In addition, the expression of important antioxidant genes, such as cat1, sod1, and glr1, was lower and non-inducible by oxidative stress in the absence of Yap1. Apart from its regulatory function in oxidative stress tolerance, Yap1 has also been attributed additional developmental and other, sometimes contradictory, roles between fungi [28]. Here, we found that in V. dahliae Yap1 is necessary for normal conidiation, in line with previous findings in M. oryzae and A. nidulans [33,37], while its deletion also severely compromised the formation of microsclerotia. A previous study demonstrated attenuated microsclerotial formation upon V. dahliae yap1 deletion [40], to a lesser extent than the one observed in our work, which probably reflects strain-specific variation in the fine-tuning of the underlying regulatory mechanisms. In agreement with the previous study that analyzed virulence on smoke trees [40], we detected no significant involvement of Yap1 in the virulence of V. dahliae on eggplant. This is consistent with studies of other fungal pathogens that do not require Yap1 for infecting their hosts (e.g., B. cinerea, Fusarium graminearum, and Cochliobolus heterostrophus), but in contrast to other fungi where Yap1 is indispensable for successful infection (e.g., Ustilago maydis, M. oryzae, and Alternaria alternata) [31,33,35,38,70,71]. These findings suggest that, in some fungi, alternative detoxification pathways or regulatory mechanisms likely take over during infection and suffice for efficient ROS detoxification when Yap1 is absent. Such alternatives could include the transcription factors Atf1 and Skn7, which mediate responses to RNS and ROS, respectively, by regulating the expression of antioxidant genes [72,73].
Based on the significant biological roles of the ROS-generating NADPH oxidase NoxA in V. dahliae, and the presumed need of the pathogen to withstand the ROS bursts involved in the defense mechanisms of its hosts [4], we would expect systems for coordination between ROS generation and processing to be important for its ROS homeostasis. In support of this hypothesis, our study revealed an antagonistic epistatic relationship between noxA and yap1, with their double null mutant exhibiting the same behavior as ∆noxA in all tested phenotypes, instead of any detectable additivity between the two genes. This was also observed with regard to sensitivity to oxidative stress, whereby ∆noxA and ∆noxA ∆yap1 were more resistant to oxidative agents than ∆yap1. Our interpretation of these findings is that Yap1 contributes to the neutralization of the ROS produced by NoxA, but this Yap1 detoxification function depends on NoxA, which itself responds to oxidative stress with higher expression levels. We propose that, in the absence of NoxA, other yap1independent systems take over, presumably via up-regulation, to effectively shield the organism from the hazards of elevated oxidative stress. This is further supported by the detected changes in the expression patterns of antioxidant genes in the ∆noxA mutant. The observed small increase in Hog1 transcripts observed in this strain is unlikely to mediate antioxidant responses, since, in V. dahliae, Hog1 appears to play no role in the OSR [74]. In other fungi, oxidative stress signaling alternatives include Pap1 activation by Tpx1 and the Hog1 homologs Sty1/Spc1 MAPK pathway, which can trigger antioxidant responses in Schizosaccharomyces pombe [75]. Furthermore, the OSR of the filamentous fungus A. nidulans involves, in addition to its Yap1 homolog NapA, the Skn7 homolog SrrA and the Hog1 homologs SakA and MpkC [76][77][78]. Finally, we would attribute the findings that low levels of oxidative stress (e.g., low concentrations of H 2 O 2 ) have no observable effect on the ∆noxA strain and, consistently, no induction of antioxidant genes, to the perturbated generation of endogenous ROS, which could be linked to a higher resistance of the fungus to low levels of exogenous oxidative stress.
In this study, we characterized the V. dahliae homolog of the important ROS producer NoxA, and we demonstrated significant pleiotropic roles in multiple developmental and physiological processes. We provide evidence that implicates NoxA in sensing and mediating responses to oxidative stress through genetic interactions with the transcriptional regulator Yap1 and presumably alternative, currently unknown, ROS signaling and detoxification systems. These data contribute to a better understanding of fungal ROS homeostasis and welcome future research to elucidate the significance and modes of action of ROS metabolism in the development, pathogenicity, and stress response of fungal pathogens.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/jof7090740/s1, Figure S1: Sequence alignment of Yap1 homologs of V. dahliae (VDAG_01588) and other fungi, Figure S2: Validation of the yap1 knockout mutants, Figure S3: Effects of antifungal agents that target the respiratory chain, and trace elements on V. dahliae deletion mutants ∆noxA and ∆yap1, Table S1: Verticillium dahliae strains constructed and used in this study, Table S2: List of plasmids constructed and used in this study, Table S3: List of DNA oligonucleotides used in this study.

Data Availability Statement:
The data that support the findings of this study are available in the main text, figures, and additional files of this article.