The thioredoxin MoTrx2 protein mediates reactive oxygen species (ROS) balance and controls pathogenicity as a target of the transcription factor MoAP1 in Magnaporthe oryzae

Summary

We have shown previously that the transcription factor MoAP1 governs the oxidative response and is important for pathogenicity in the rice blast fungus Magnaporthe oryzae. To explore the underlying mechanism, we have identified thioredoxin MoTrx2 as a target of MoAP1 in M. oryzae. Thioredoxins are highly conserved 12‐kDa oxidoreductase enzymes containing a dithiol–disulfide active site, and function as antioxidants against free radicals, such as reactive oxygen species (ROS). In yeast and fungi, thioredoxins are important for oxidative stress tolerance and growth. To study the functions of MoTrx2, we generated ΔMotrx2 mutants that exhibit various defects, including sulfite assimilation, asexual and sexual differentiation, infectious hyphal growth and pathogenicity. We found that ΔMotrx2 mutants possess a defect in the scavenging of ROS during host cell invasion and in the active suppression of the rice defence response. We also found that ΔMotrx2 mutants display higher intracellular ROS levels during conidial germination, but lower peroxidase and laccase activities, which contribute to the attenuation in virulence. Given that the function of MoTrx2 overlaps that of MoAP1 in the stress response and pathogenicity, our findings further indicate that MoTrx2 is a key thioredoxin protein whose function is subjected to transcriptional regulation by MoAP1 in M. oryzae.

Introduction

Rice blast, inducing the most serious economic losses of rice worldwide, is caused by the filamentous fungus Magnaporthe oryzae. The pyriform conidia attach and germinate on the hydrophobic plant surface, before forming a dome‐shaped appressorium at the end of the germ tube (Talbot, 2003; Wilson and Talbot, 2009; Zhang et al., 2016a). Mature appressoria accumulate osmolytes to generate the enormous turgor pressure (8 MPa) required to penetrate the plant cuticle and to invade the host (Howard et al., 1991). The hyphae ramify through the host cells, resulting in disease lesions that can produce numerous conidia in high humidity. Subsequently, these conidia can spread rapidly by wind or rain to initiate new disease cycles (Talbot, 2003).

In response to pathogenic attack, plants display innate defence systems (Chi et al., 2009), such as the rapid accumulation of reactive oxygen species (ROS) at the site of pathogen invasion. This is often considered as the first line of defence against invading pathogens (Apostol et al., 1989; Chi et al., 2009). Host ROS accumulation plays a role in antimicrobial defence by either strengthening the cell wall through oxidative cross‐linking to obstruct the expansion of the pathogen or by triggering programmed cell death to block pathogen invasions (Apel and Hirt, 2004; Guo et al., 2010). In plant–microbe interactions, plant cells initiate ROS by activating plasma membrane NADPH oxidases, which lead to the cellular defence against pathogens (Lamb and Dixon, 1997; Yoshioka et al., 2003). The interaction between the small GTPase Rac and the gp91phox subunit Rboh is also thought to be important for the activation of the NADPH oxidase in rice (Ono et al., 2001; Wong et al., 2007). As the oxidative burst reaction occurs rapidly, the pathogen must quickly counteract this stress. Plant pathogens have evolved various strategies to rapidly detoxify ROS through cellular enzymatic and non‐enzymatic mechanisms. Many of these rapid oxidative stress responses are also regulated through stress‐responsive transcription factors (Apel and Hirt, 2004; Guo et al., 2010, 2011; Kim et al., 2009; Liu et al., 2010; Qi et al., 2012; Tang et al., 2015). We have shown previously that the transcription factor MoAP1 is a key factor in the regulation of gene expression involved in the detoxification of oxidative stress by regulating laccase and peroxidase activities (Guo et al., 2011).

Thioredoxins are ubiquitous, small, heat‐stable oxidoreductases that maintain the redox homeostasis of cells (Aguirre et al., 2005; Holmgren, 1989; Trotter and Grant, 2002; Vignols et al., 2005). There are two cytoplasmic thioredoxins, Trx1 and Trx2, found in the budding yeast Saccharomyces cerevisiae that participate in the redox state balance, DNA replication and function of 3′‐phosphoadenosine 5′‐phosphosulfate (PAPS) reductase (Muller, 1991; Trotter and Grant, 2002). In the plant Arabidopsis thaliana, Trx h9 and Trx h3 are the members of the Arabidopsis Trx h family that maintain the cellular redox balance (Meng et al., 2010).

A recent study has found that there are two thioredoxins, MoTrx1 and MoTrx2, in M. oryzae, and that MoTrx2 interacts with MoMst7 of the PMK1 mitogen‐activated protein (MAP) kinase pathway to control appressorium function and infection (Zhang et al., 2016b). Here, we provide evidence to show how MoTrx2 may exert its function through regulation by the transcription factor MoAP1. We also reveal additional functions of MoTrx2, in particular its roles in mediating the oxidative stress response and pathogen–host interaction.

Results

Identification of MoTrx2 as a target of MoAP1

We have found previously that the transcription factor MoAP1 regulates the genes involved in morphogenesis, the ROS balance and pathogenicity in M. oryzae (Guo et al., 2011). MoTRX2, encoding MoTrx2, is amongst those mostly affected and its expression is decreased significantly in the ΔMoap1 mutant (Fig. 1A). To study how MoAP1 might regulate MoTrx2, we first identified the putative MoAP1 binding sites in the promoter region of MoTRX2 by performing an electrophoretic mobility shift assay (EMSA). The Alex660‐labelled DNA containing the promoter sequence was retarded by the addition of the purified MoAP1 protein, and this retardation increased drastically as the amount of MoAP1 increased (Fig. 1B,C). In addition, this DNA band shift was diminished when proteinase K was added (Fig. 1B). To further examine the binding of MoAP1, the MoTRX2 promoter was divided into five regions and each region was incubated with MoAP1. The results showed that the MoTRX2 promoter contains a MoAP1 binding site located 500 bp upstream of the start codon (Fig. 1D).

Figure 1.

Electrophoretic mobility shift assay of MoAP1 binding to the MoTRX2 promoter. (A) Expression analysis of MoTRX2 in the ΔMoap1 mutant and wild‐type. (B) The Alex660‐labelled full‐length DNA of the promoter was incubated in the absence (leftmost lane) or presence of purified MoAP1 (second lane) and β‐glucuronidase (GST) protein (fourth lane). Proteinase K was added after the incubation of MoAP1 with the DNA (third lane). DNA–protein complexes were separated by electrophoresis on a 1% agarose gel. (C) Increasing amounts of MoAP1 were incubated with Alex660‐labelled 1500‐bp DNA. The complexes were resolved by electrophoresis on a 1% agarose gel. (D) The promoter of MoTRX2 was divided into five parts for incubation with MoAP1 to examine the binding site.

We also examined the expression of MoTRX2 in M. oryzae, and found that it was much higher in conidia and infection stages than in the mycelial stage (Fig. S1, see Supporting Information), suggesting that MoTRX2 may play a potential role in sporulation and infection in M. oryzae. Interestingly, although the expression of MoTRX2 was decreased in the ΔMoap1 mutant, overexpression of MoTRX2 had no apparent effect in restoring the pathogenicity defects of the ΔMoap1 mutant (Fig. S2, see Supporting Information).

MoTrx2 plays a critical role in vegetative growth and sulfite assimilation

In S. cerevisiae, thioredoxins participate in the reduction process of PAPS to sulfite ( ${\text{SO}}_3^{2–}$). Given that yeast Trx2 was identified as relevant to sulfate assimilation prior to methionine metabolism (Thomas and Surdin‐Kerjan, 1997), we investigated whether MoTrx2 played a similar role in M. oryzae. Two mutants (#4 and #9) were obtained by replacing the MoTRX2 coding region with the hygromycin resistance cassette (HPH) and verified by Southern blot analysis (Fig. S3, see Supporting Information). When incubated on complete medium (CM) and minimal medium (MM) at 28 ºC for 7 days in the dark, ΔMotrx2 mutants showed reduced growth on both media, with no aerial hyphae on MM (Fig. 2A, Table S1, see Supporting Information), indicating that MoTrx2 might have a role in mineral salt utilization. As MM contains ${\text{SO}}_4^{2–}$ and auxotrophy indicates that ΔMotrx2 mutants have defects in ${\text{SO}}_4^{2–}$ reduction, we wondered whether supplementation of exogenous ${\text{SO}}_3^{2–}$ and S^2– could compensate for the defect seen in ΔMotrx2 mutants. Indeed, this defect was complemented by the addition of exogenous ${\text{SO}}_3^{2–}$ and S^2–, albeit partially (Fig. 2B), further indicating that MoTrx2 is involved in sulfite assimilation.

Figure 2.

ΔMotrx2 mutants show defects in growth and sulfate assimilation.(A) Guy11, ΔMotrx2 mutants and complemented strain were inoculated on complete medium (CM) and minimal medium (MM), cultured at 28 °C for 7 days and then photographed. (B) Diagram of sulfate assimilation in fungi and growth rate of ΔMotrx2 mutants on MM plates with sulfate. Guy11, mutants and complement strain were inoculated on MM plates supplemented with 1.25 mm sulfate ( ${\text{SO}}_3^{2–}$) and 1.25 mm sulfite (S^2–). The strains were cultured at 28 ºC for 5 days. ASP, adenylyl sulfate; MET16, 3′‐phosphoadenosine 5′‐phosphosulfate reductase; PAPS, 3′‐phosphoadenosine 5′‐phosphosulfate.

MoTrx2 is required for asexual and sexual development

We found that there were no significant differences in conidium morphology and germination between the ΔMotrx2 mutants and the wild‐type strain Guy11 (Table S1). However, the number of conidia was significantly decreased in the ΔMotrx2 mutant (Fig. S4A, see Supporting Information), similar to that reported by Zhang et al. (2016b). To examine the role of MoTrx2 in sexual reproduction, the wild‐type, ΔMotrx2 mutants and complemented strains (MAT1‐2) were tested in confrontation with the standard test strain TH3 (MAT1‐1). Perithecia were observed at the junctions of Guy11 and complemented strains, but not ΔMotrx2 mutants, with TH3 (Fig. S4B, top panel) after 3 weeks. Consistent with this observation, microscopic examination revealed abundant mature asci and ascospores between Guy11 and TH3, but no ascus or ascospore between ΔMotrx2 and TH3 (Fig. S4B, bottom panel). The results indicate that MoTrx2 plays an essential role in mating.

MoTrx2 is important for pathogenicity and infectious hyphal growth

To determine whether MoTrx2 is involved in pathogenesis, conidial suspensions (5 × 10⁴ conidia/mL), harvested from wild‐type, two independently generated ΔMotrx2 mutants and complemented strains, were sprayed onto susceptible rice seedlings (CO‐39 cv. oryzae). After 7 days, the ΔMotrx2 mutants caused disease lesions with a significant reduction in lesion expansion compared with the wild‐type (Fig. 3A). In addition, the restricted lesions did not produce conidia under constant illumination (Fig. 3B). To quantify the disease on rice, we used a ‘lesion‐type’ scoring assay (Wang et al., 2013). The assay showed that lesions produced by the ΔMotrx2 mutants were less severe, with most rated as type 1–2 and fewer as type 3–4 (Fig. 3C). Similar results were also observed in the rice injection assay (Fig. 3D).

Figure 3.

ΔMotrx2 mutants show defects in pathogenesis. (A) Two‐week‐old rice seedlings (Oryza sativa cv. CO‐39) were sprayed with conidial suspension (1 × 10⁵ conidia/mL). The diseased leaves were harvested at 7 days. Three independent experiments were performed. (B) The infected rice leaves were illuminated for 24 h to produce conidia. The arrows point to conidia. The lesions were observed under a light microscope. (C) Lesion type statistics (0, no lesion; 1, dark‐brown pinpoint lesions; 2, 1.5‐mm brown spots; 3, 2–3‐mm lesions with brown margins; 4, eyespot lesions longer than 3 mm; 5, coalesced lesions infecting 50% or more of the leaf maximum size). Asterisks represent significant difference (P < 0.01). (D) Conidial suspension (1 × 10⁵ conidia/mL) was injected in rice sheaths of 3‐week‐old seedlings, and 60 healthy rice seedlings were used for each strain.

As no defects in appressorial turgor pressure were found in the ΔMotrx2 mutants (Table S1), we inoculated conidia on barley epidermis and rice sheaths to examine the invasion hyphae. We found that 70% of invasion hyphae were type 4 in barley infected with Guy11 after 24 h, whereas less than 10% and 60% of invasion hyphae were type 4 and type 3, respectively, in barley infected with ΔMotrx2 mutants (Fig. S5A, see Supporting Information). In rice sheath cells, the wild‐type invasion hyphae expanded to neighbouring cells at 48 h post‐inoculation (hpi); however, the ΔMotrx2 mutants were mostly restricted within the primary infection sites (Fig. S5B). These findings are again in accordance with those described by Zhang and colleagues (Zhang et al., 2016b), suggesting that the attenuated virulence of ΔMotrx2 mutants is caused by a defect in infectious hyphal growth.

MoTrx2 plays a critical role in the scavenging of ROS and induction of the plant defence response

In plants, the generation of ROS is one of the first responses to fungal invasion (Chi et al., 2009). To successfully invade host cells and adapt to the changing environment, pathogens must develop mechanisms to protect against ROS‐induced damage (Garrido and Grant, 2002; Guo et al., 2010, 2011; Liu et al., 2016). To test whether MoTRX2 is important for the scavenging of ROS in the plant cell, ROS production was assayed with 3,3′‐diaminobenzidine (DAB) staining of rice sheaths following infection. ROS was barely detectable in rice cells infected with the wild‐type strain; however, an enhanced amount of ROS was detected in host cells infected with the ΔMotrx2 mutant (Fig. 4A,B). To test the hypothesis that the deficiency of hyphal expansion in ΔMotrx2 mutants was caused by a defect in the scavenging of host ROS, we used diphenyleneiodonium (DPI) to inhibit the activity of plant NADPH oxidases that are necessary for ROS generation in plants (Chen et al., 2014). After incubation at 28 ºC for 36 hpi, one layer of rice sheath endodermis cells treated with 0.5 µM DPI was observed under a light microscope. Approximately 70% of mutants formed type 3 and type 4 infectious hyphae. Without DPI, the infectious hyphae of the ΔMotrx2 mutants in plant cells had just one or two branches (Fig. 4A,C). This observation was in agreement with our hypothesis.

Figure 4.

MoTrx2 participates in the scavenging of reactive oxygen species (ROS). (A) Conidial suspensions of four different strains were injected in separate rice sheaths. At 36 h post‐inoculation (hpi), 3,3′‐diaminobenzidine (DAB) was used to stain the sheaths for 8 h. Conidial suspensions treated with or without 0.5 µM diphenyleneiodonium (DPI) dissolved in dimethylsulfoxide (DMSO) were infected in rice sheaths; the results were observed at 36 hpi. Bar, 20 µm. (B) Percentages of cells with infectious hyphae were dyed by DAB. Means were calculated from three independent replicates. Significant differences are indicated by asterisks: P < 0.01; n = 100. (C) Percentages of different types of infectious hyphae in rice cells treated with DPI or DMSO. The rice sheath treated with DMSO was a negative control. Three independent experiments were replicated.

As a result of the innate defence system, most plants exhibit immunity to pathogen infections. However, M. oryzae can cause diseases by evading rice recognition and proliferating in live host cells through the active suppression of plant immunity (Chi et al., 2009; Mentlak et al., 2012). In addition, host immune systems result in the reinforcement of the cell wall and the accumulation of pathogenesis‐related (PR) proteins in infected cells (Chen et al., 2014). To further determine whether the deletion of MoTRX2 could activate plant defence gene expression, we examined the transcript levels of PBZ1, AOS2, PR1a and CHT1 genes, which are involved in jasmonic acid (JA) or salicylic acid (SA) pathways (Dong et al., 2015; Guo et al., 2010), using real‐time quantitative polymerase chain reaction (PCR). The expression of PR1a, PBZ1 and AOS2 was markedly higher at 72 than at 24 hpi in rice infected with the ΔMotrx2 mutant (Fig. 5). Taken together, we propose that the restricted expansion of the ΔMotrx2 mutant in rice cells is, at least in part, caused by the defect in active suppression of the rice defence response.

Figure 5.

ΔMotrx2 mutant induces defence‐responsive gene expression in infection stages. The transcription of PR1a, AOS2, PBZ1 and CHT1 in the infected host was assayed using quantitative reverse transcription‐polymerase chain reaction (qRT‐PCR). The RNA samples were collected from infected rice leaves at 8, 24 and 72 h post‐inoculation (hpi). The average threshold cycle of triplicate reactions was normalized by the stable‐expression gene elongation factor 1a (Os03g08020) in Oryza sativa. The assay was repeated three times with similar results obtained.

MoTrx2 is important for intracellular ROS balance during conidial germination and the early stage of appressorial formation

Previous studies have indicated that an alteration in intracellular ROS affects virulence in rice seedlings by M. oryzae (Zhang et al., 2010), and that the NADPH oxidase mutants lose virulence on susceptible rice cultivar CO‐39 (Egan et al., 2007). To assess ROS levels, we used dihydrorhodamine 123, which exhibits green fluorescence during reduction by superoxide radicals. Green fluorescence was more intense in the germ tubes and early appressorial formation of the ΔMotrx2 mutants than in the wild‐type or the complemented strain (Fig. 6A). Quantitative measurement of the relative fluorescence density indicated that the most apparent difference occurred at 2 h of appressorial formation (Fig. 6B). To confirm this, we carried out an additional assay using nitroblue tetrazolium (NBT), which forms a dark‐blue water‐insoluble formazan on reduction by superoxide radicals (Song et al., 2010; Tanaka et al., 2006). Microscopic observation revealed that the superoxide content was increased in ΔMotrx2 mutants in comparison with the wild‐type strain, with more formazan found in immature appressoria at 2 h (Fig. 6C). Thus, both staining assays indicated that MoTRX2 deletion leads to an increased accumulation of superoxide during conidial germination and the early stages of appressorial formation.

Figure 6.

Accumulation of intracellular reactive oxygen species (ROS) increases in the ΔMotrx2 mutant during infection. (A) Dihydrorhodamine 123 was used to detect superoxide during conidial to appressorial formation. The conidia were stained after inoculation on hydrophobic cover slips for 2, 4 and 24 h. Before being viewed by epifluorescence microscopy, the dye was rinsed twice with phosphate‐buffered saline (PBS). Fluorescence images were captured by a 100‐ms exposure for absorbed light using a green fluorescent protein (GFP) filter. (B) Relative fluorescence density calculation. Significant differences are indicated by double asterisks: P < 0.01. (C) Detection of superoxide by nitroblue tetrazolium (NBT) staining. Samples were prepared as above, stained with NBT aqueous solution for 1 h and viewed by light microscopy at each time point after rinsing with water. The experiment was repeated three times with similar results.

ΔMotrx2 mutant exhibits lower extracellular peroxidase and laccase activity

The activities of extracellular peroxidase and laccase are important in the fungal detoxification of host‐derived ROS during infectious hyphal extension and colonization (Chi et al., 2009; Song et al., 2010). This can be assayed using 2,2′‐azino‐bis(3‐ethylbenzothiazoline‐6‐sulfonic acid) diammonium salt (ABTS) as the substrate. We found that both laccase and extracellular peroxidase activities were decreased in ΔMotrx2 mutants (Fig. 7A). The decreased activities may be caused by reduced transcription of these enzyme encoding genes (Guo et al., 2010). We thus detected the expression of several genes encoding extracellular peroxidase and laccase, and found that most of the transcripts were decreased in the ΔMotrx2 mutant (Fig. 7B,C).

Figure 7.

MoTrx2 plays a role in the regulation of extracellular laccase and peroxidase activities. (A) Laccase and peroxidase activities measured by the 2,2′‐azino‐bis(3‐ethylbenzothiazoline‐6‐sulfonic acid) diammonium salt (ABTS) oxidizing test without or with H₂O₂ treatment. (B, C) Extracellular laccase and peroxidase gene expression levels were altered in the ΔMotrx2 mutant. The expression profiles of two laccases and nine putative peroxidases in the Guy11 and mutant strains. The experiments were repeated three times. Asterisks indicate a significant difference: P < 0.01.

Disulfide bond formation is essential for the function of MoTrx2

Previous studies by others have shown that thioredoxins interact with other proteins through the cysteine (Cys) at the N‐terminus, which contains a conservative sequence Cys–X–X–Cys (Calvo et al., 2013; Vignols et al., 2005). MoTrx2 contains three cysteines, C3, C95 and C98, with C95 and C98 being conserved within the Cys–Gly–Pro–Cys active site (Holmgren, 1989). To further dissect MoTrx2 function, we performed alanine substitution studies. Three site‐directed mutation strains (ΔMotrx2/MoTRX2 ^C3A, ΔMotrx2/MoTRX2 ^C95A and ΔMotrx2/MoTRX2 ^C98A) were obtained and the phenotypes were evaluated. Colony growth on MM, sporulation and pathogenicity of ΔMotrx2/MoTRX2 ^C95A and ΔMotrx2/MoTRX2 ^C98A were similar to those of the ΔMotrx2 mutants, except ΔMotrx2/MoTRX2 ^C3A (Fig. 8). These results were consistent with the proposition that the disulfide bond exists between C95 and C98 and that it plays an essential role in MoTrx2 function.

Figure 8.

Disulfide bond formation in MoTrx2 is important for growth, conidiation and pathogenesis. (A) Guy11, mutants and point mutation strains were inoculated on minimal medium (MM) plates supplemented with sulfate ( ${\text{SO}}_3^{2–}$) and sulfite (S^2–). The strains were cultured at 28 ºC for 5 days before being photographed. (B) Rice seedlings were sprayed with conidial suspension, keeping warmth and moisture for 7 days. Three independent experiments were performed. (C, D) Conidial formation was observed under a light microscope after illumination for 24 h. The conidia were collected, counted and analysed by Duncan analysis (P < 0.01). Asterisks indicate significant differences and the results were repeated three times.

Discussion

In fungi, the thioredoxin proteins are closely linked to various cellular differentiation and developmental processes. Saccharomyces cerevisiae Trx2 functions in sulfur metabolism, cell cycle regulation and resistance to oxidative stress (Garrido and Grant, 2002; Muller, 1991). In Saccharomyces pombe, Trx2 is required for Pap1 oxidation, whereas the peroxiredoxin Tpx1 protein regulates Pap1 independent of Trx1 (Brown et al., 2013; Calvo et al., 2013; Day et al., 2012). In Aspergillus nidulans, thioredoxins are important in redox regulation and cellular development (Wong et al., 2007). Although MoTrx2 was first described by Zhang and colleagues (Zhang et al., 2016b), our studies revealed additional functions of MoTrx2. Moreover, we linked MoTrx2 to the function of the transcription factor MoAP1.

By employing EMSA, we demonstrated that MoAP1 binds the MoTRX2 promoter region. Surveying the predicted AP1 binding site motif in yeast (regions containing MTTACGTAAK, TTAGTMAGC and TTASTMA), we identified the most conserved regions in the promoter of MoTRX2 in M. oryzae. However, we found that these predicted binding sites did not affect MoAP1 binding, but, instead, the binding site was 500 bp upstream of the transcription initiation sequence. The expression level of MoTRX2 was severely decreased in the Moap1 mutant strain, suggesting a functional correlation. However, overexpression of MoTRX2 had no effect on the pathogenicity of the ΔMoap1 mutant on both rice and barley. Combined with our previous findings that MoAP1 regulates the expression levels of multiple genes related to pathogenicity, including MoGTI1, MoPAC2 and MoYCP4 (Chen et al., 2014, 2016), MoTRX2 expression alone is not sufficient to restore or compensate for the functions of MoAP1.

Eukaryotic microorganisms use the sulfur atom for biosynthesis and most are able to perform the assimilatory reduction of sulfate (Thomas and Surdin‐Kerjan, 1997). In S. cerevisiae, the thioredoxin and PAPS reductase (MET16) form a homodimer to play a role in sulfate reduction (Berendt et al., 1995). In the prokaryote Escherichia coli, thioredoxin and glutaredoxin participate in the reduction of sulfate to sulfite by PAPS reductase (Russel et al., 1990). When MM was supplemented with ${\text{SO}}_3^{2–}$ and S^2–, the ΔMotrx2 mutants could partially suppress the defects of aerial hyphal growth. As the ΔMotrx2/MoTRX2^C95A and ΔMotrx2/MoTRX2^C98A mutants showed the same phenotypes as ΔMotrx2 mutants on MM, this indicates that the disulfide bond of the CGPC motif participates in the reduction of sulfate to sulfide. On MM, the phenotypes of the ΔMomet16 mutant were in line with those of the ΔMotrx2 mutants. As an interaction between MoTrx2 and MoMet16 cannot be established through yeast two‐hybrid assay and MoMet16 has no effect on pathogenicity (Fig. S6, see Supporting Information), we propose that MoTrx2 may be involved in pathogenicity through mechanisms other than the sulfur source assimilation pathway. In addition, ${\text{SO}}_4^{2–}$, ${\text{SO}}_3^{2–}$ and S^2– were added to straw decoction and corn (SDC) agar to identify the relationships between the defect of sulfate assimilation and reduced conidial formation; however, none could rescue the defect in conidiation. These results indicate that the reduction in conidial formation of the ΔMotrx2 mutant is independent of sulfate assimilation. We thus conclude that MoTrx2 plays an important role in inorganic sulfur source assimilation, which is consistent with previous results found in other fungi.

In this study, we observed that the ΔMotrx2 mutants exhibited a dramatic defect in virulence. This could result from multiple defects exhibited by the mutants. First, intracellular ROS is increased during appressorial formation. Intracellular ROS production is a ubiquitous eukaryotic signalling system for the control of various fungal functions, including conidia germination, sexual development, immunity and apoptosis, in A. nidulans, Podospora anserine and Neurospora crassa (Aguirre et al., 2005; Malagnac et al., 2004; Takemoto et al., 2006). In addition, deletion of Alternaria brassicicola tmpL induced an accumulation of intracellular ROS in conidia and infection structures, and the tmpL mutant exhibited reduced virulence on green cabbage leaves (Kim KH et al., 2009). In M. oryzae, MoNox1 and MoNox2 NADPH oxidases are the potential source of ROS production and ROS is important for plant infection (Egan et al., 2007). In our study, the comparison of intracellular ROS between the ΔMotrx2 mutant and the wild‐type strain revealed that more ROS was accumulated during the early stage of appressorial formation in the Motrx2 mutant. We speculated that MoTrx2 participates in balancing intracellular ROS homoeostasis during conidial germination, and the imbalance of intracellular ROS in ΔMotrx2 mutants is one of the reasons for the attenuated pathogenicity.

Second, the ΔMotrx2 mutants showed a defect in scavenging host ROS during infection that could induce a potent plant defence response. Host plants can accumulate ROS at the infection site as a response during plant immunity via the pathogen‐associated molecular pattern (PAMP)‐triggered immunity (PTI) pathway to restrain pathogen penetration and colonization. In response to ROS, pathogens develop mechanisms to scavenge exogenous ROS for survival and invasion of the host cell (Molina and Kahmann, 2007; Tanaka et al., 2006). Fungi secrete peroxides to help pathogens detoxify host‐derived ROS during plant–microbe interactions (Guo et al., 2010; Molina and Kahmann, 2007). In M. oryzae, MoDes1, MoAtf1 and MoAP1 regulate the expression of peroxidase and laccase, and are required for pathogenicity. As the ΔMotrx2 mutants showed restricted infectious hyphal growth and reduced virulence (Fig. 4A), we used DAB staining assay to demonstrate that the ΔMotrx2 mutants showed a defect in the scavenging of ROS; we also showed that the infection defect of mutants can be partially suppressed when adding DPI to suppress plant ROS production.

PTI can induce plant defence gene expression. Plants possess salicylic acid (SA) dependent resistance pathway and jasmonic acid (JA) resistance pathway, which synergistically or antagonistically defend against pathogen invasion (Dong et al., 2015; Qiu et al., 2007). Many of the defence‐related genes are involved in the JA and SA pathways, such as PR1a, Cht1, PBZ1, AOS2, Lox and PAD4 (Qiu et al., 2007). The quantitative reverse transcription‐polymerase chain reaction (qRT‐PCR) data showed that the expression levels of these defence‐related genes were all up‐regulated (except for CHT1 at 8 hpi) in the ΔMotrx2 mutant compared with that in the Guy11 strain, which was consistent with the finding that the accumulation of ROS can induce the expression of plant defence‐related genes.

Taken together, our studies demonstrate that MoTrx2, a member of the thioredoxin family, is important for conidiation, sexual reproduction and pathogenicity. We found that MoTrx2 regulates the expression of extracellular peroxidase and laccase genes, and induces plant defence responses. Thioredoxins play a role in the intracellular ROS balance necessary for full virulence. Importantly, we showed that MoAP1 function is integral to MoTrx2 function, and MoTrx2 is subject to transcriptional regulation by MoAP1. Given the important roles of thioredoxins in rice blast pathogenicity, as demonstrated by us and Zhang et al. (2016b), and the urgency to develop novel rice blast control strategies, further studies on thioredoxin proteins are warranted.

Experimental Procedures

Strains and culture conditions

The M. oryzae Guy11 strain was used as wild‐type for transformation in this study. All strains were cultured on CM for 7 days at 28 ºC (Zhang et al., 2011). Liquid medium was used to prepare the mycelia for genomic DNA and RNA extraction. For conidial collection, mycelial blocks were inoculated on SDC (100 g of straw, 40 g of corn powder, 15 g of agar in 1 L of distilled water) agar medium (Zhang et al., 2011), maintained at 28 °C for 7 days in the dark, followed by 3 days of continuous illumination under fluorescent light.

MoTRX2 deletion and mutant complementation

Using a one‐step gene replacement strategy, MoTRX2 was replaced with the hygromycin resistance cassette (HPH). First, two 1.0‐kb sequences flanking the targeted gene were PCR amplified with primer pairs (Table S2, see Supporting Information). Then, the PCR products were digested with restriction endonucleases and ligated with pCX62, which contains a hygromycin resistance cassette. Two flanking sequences were added to different sides of HPH forming a 3.4‐kb fragment. The fragments were amplified and transformed into protoplasts of Guy11. Transformants were screened by PCR and confirmed by southern blotting analysis (Fig. S3). The complement fragments contain the MoTRX2 gene and its native promoter region; the fragments were amplified by PCR with primers and inserted into pYF11 (bleomycin resistance) to complement the ΔMotrx2 mutant.

Assays for vegetative growth

For hyphal growth, small agar blocks were cut from the edge of 4‐day‐old cultures and placed onto fresh media (CM and MM) in the dark at 28 ºC for 7 days (Zhang et al., 2010); diameters were measured and photographed. To test hyphal growth on different sulfur sources, strains were cultured on MM [1% glucose MM: 6 g/L NaNO₃, 0.52 g/L KCl, 0.52 g/L MgSO₄.7H₂O, 1.52 g/L KH₂PO₄, 10 g/L glucose, 0.001% (w/v) thiamine and 0.1% (w/v) trace elements]. For inorganic sulfur sources, the same amount of sulfate ( ${\text{SO}}_4^{2–}$), sulfite ( ${\text{SO}}_3^{2–}$) and sulfide (S^2–) were added to MM. All experiments were repeated three times, each with three replicates.

Conidiation, appressorial formation and sexual reproduction

Conidia were harvested from 10‐day‐old SDC cultures, filtered through three layers of lens paper and counted under microscopy. For appressorial formation, droplets (30 µL) of conidial suspension (5 × 10⁴ spores/mL) were placed on plastic coverslips (hydrophobic) incubated under humid conditions at room temperature, as described previously (Zhang et al., 2010). Appressorial formation was examined after incubation for 24 h. All experiments were repeated three times, each with three replicates.

Sexual reproduction assays were performed on Guy11 (MAT1‐2) and the tested strains with the standard tester strain TH3 (MAT1‐1) on oatmeal medium (OM) plates. Incubation was performed at 28 ºC before the two colonies were contacted, followed by 4 weeks under continuous white fluorescent light at 20 ºC (Zhang et al., 2011).

Pathogenicity assay

For spray inoculation, conidia were resuspended to a concentration of 5 × 10⁴ spores/mL in a 0.2% (w/v) gelatine solution. A conidial suspension was sprayed onto 2‐week‐old seedlings of rice (Oryza sativa cv. CO39) and 7‐day‐old barley. Inoculated plants were kept in a growth chamber at 25 ºC with 90% humidity and in the dark for the first 24 h, followed by a 12 h/12 h light/dark cycle for 7 days. For microscopic observation of hyphal expansion in plant cells, ten 10 µL droplets with the same concentration of conidia were placed onto the lower epidermis of detached barley leaves, maintained in the dark for 24 h, and then observed and photographed under light microscopy. Detached rice sheaths were injected with 5 × 10⁴ spores/mL conidia, followed by incubation at 28 ºC for 36 h. All experiments were repeated three times, each with three replicates.

ROS detection assay

ROS was observed by staining with DAB (Sigma‐Aldrich, St. Louis, MO, USA) after rice sheaths had been inoculated with conidia for 36 hpi. Infected sheaths were incubated in 1 mg/mL DAB solution for 8 h and destained with clearing solution (ethanol acetic acid, 94 : 4 v/v) for 1 h (Chen et al., 2014). The conidial suspension supplemented with 0.5 µm DPI was infected in rice sheaths to inhibit plant ROS generation. Intracellular ROS levels were measured during appressorial formation using dihydrorhodamine 123 (Invitrogen, Carlsbad, California, USA) and NBT as oxidant‐sensitive probe. The appressorium at different times was stained by 50 µM dihydrorhodamine 123 and 0.3 mM NBT aqueous solution as described previously (Chen et al., 2014; Guo et al., 2010). The results were viewed by epifluorescence microscopy and both dihydrorhodamine 123 and NBT staining were performed three times.

Measurement of extracellular peroxidase and laccase activities

Mycelia were cultured in liquid CM for 2 days. The culture filtrate was collected, and peroxidase and laccase activities were measured using colorimetric determination (Chi et al., 2009). Three independent biological experiments were performed.

qRT‐PCR and gene expression analysis

For qRT‐PCR, total RNAs were reverse transcribed into first‐strand cDNA using the Reverse Transcription Kit (Vazyme, Beijing, China). qRT‐PCR was run on an Applied Biosystems (Foster City, CA, USA) 7500 Real Time PCR System with SYBR Premix ExTaq (Perfect Real Time, Takara, Japan). The experiment was conducted twice with three independent biological replicates.

EMSA

The MoAP1 protein was expressed and purified from E. coli strain Rosetta using the pGEX4T‐2 construct containing an N‐terminal β‐glucuronidase (GST)‐tag coding sequence. The DNA fragment from the MoTRX2 promoter was end‐labelled with Alex660 by PCR amplification using the 5′ Alex660‐labelled primer. The purified protein was mixed with Alex660‐labelled DNA, incubated for 20 min at 25 °C in binding buffer and separated by agarose gel electrophoresis. Gels were visualized directly using a LI‐COR (Lincoln, NE, USA) Odyssey scanner with excitation at 700 nm.

Supporting information

Additional Supporting Information may be found in the online version of this article at the publisher's website:

Fig. S1 The expression of MoTRX2 in different stages of fungal development. The phase‐specific expression of MoTRX2 was quantified by quantitative reverse transcription‐polymerase chain reaction (qRT‐PCR) with the synthesis of cDNA from each sample, including infectious growth, vegetative growth and conidia. CO, conidia; hpi, h post‐inoculation; HY, hyphae.

Fig. S2 Overexpression of MoTRX2 has no effect on pathogenicity in ΔMoap1. (A) The expression levels of MoTRX2 in Guy11, ΔMoap1 mutant and ΔMoap1/MoTRX2 strains. (B, C) Pathogenicity test on rice and barley. Detached rice leaves were inoculated with mycelium from Guy11, ΔMoap1 and ΔMoap1/MoTRX2 mutants and photographed after 5 days. The experiments were performed three times with similar results.

Fig. S3 MoTRX2‐targeted gene knockout strategy. (A) A 0.65‐kb fragment of the MoTRX2 coding region was replaced with a 1.4‐kb hygromycin resistance cassette (HPH). Arrows indicate target gene and HPH gene. Thin lines below the arrows indicate the probe sequence to validate MoTRX2 deletion. (B) Southern blot analysis was used to identify the MoTRX2 deletion and the copy of the HPH gene. The genomic DNA of Guy11 and ΔMotrx2 mutants was digested with HindIII and hybridized with probes.

Fig. S4 MoTrx2 is required for conidiation and sexual reproduction. (A) Conidia formation was observed under a light microscope after black light lamp irradiation for 24 h. (B) Perithecia development and ascospores were observed after inoculation for 3 weeks. Cross between TH3 (MAT1‐1) and Guy11 (MAT1‐2) represents the positive control. Arrows indicate perithecia. Bar, 10 µm.

Fig. S5. MoTrx2 is important for infectious hyphal growth. (A) Quantification of infectious hyphae on barley epidermis. Infectious growth was divided into four types; the percentages of the different types are shown. Three independent experiments were performed. n = 100. Bar, 20 µm. (B) Microscopic observation of infectious hyphal expansion on rice leaf sheath cells. In vitro rice sheaths from 4‐week‐old rice were inoculated with spore suspension and examined at 48 h post‐inoculation (hpi). Bar, 20 µm.

Fig. S6 MoMet16 shows a defect in sulfate assimilation, but no defect in infection. (A) Guy11, ΔMomet16 mutant and complementary strains were inoculated on minimal medium (MM) plates supplemented with sulfate ( ${\text{SO}}_3^{2–}$) and sulfite (S^2–). The strains were cultured at 28 ºC for 5 days. (B) Rice seedlings (Oryza sativa cv. CO‐39) were sprayed with conidial suspension (4 mL, 1 × 10⁵ conidia/mL) and photographed after 7 days. Three independent experiments were performed. (C) Yeast transformants expressing bait (pGBKT7) and prey (pGADT7) constructs were assayed for growth on SD–Leu–Trp–His–Ade and SD–Leu–Trp plates and for β‐galactosidase (LacZ) activities with positive and negative control.

Table S1 Comparison of mycological characteristics among strains.

Table S2 Primers used in this study.