Auxilin‐like protein MoSwa2 promotes effector secretion and virulence as a clathrin uncoating factor in the rice blast fungus Magnaporthe o ryzae

Summary

• Plant pathogens exploit the extracellular matrix (ECM) to inhibit host immunity during their interactions with the host. The formation of ECM involves a series of continuous steps of vesicular transport events.

• To understand how such vesicle trafficking impacts ECM and virulence in the rice blast fungus Magnaporthe oryzae, we characterised MoSwa2, a previously identified actin‐regulating kinase MoArk1 interacting protein, as an orthologue of the auxilin‐like clathrin uncoating factor Swa2 of the budding yeast Saccharomyces cerevisiae.

• We found that MoSwa2 functions as an uncoating factor of the coat protein complex II (COPII) via an interaction with the COPII subunit MoSec24‐2. Loss of MoSwa2 led to a deficiency in the secretion of extracellular proteins, resulting in both restricted growth of invasive hyphae and reduced inhibition of host immunity. Additionally, extracellular fluid (ECF) proteome analysis revealed that MoSwa2‐regulated extracellular proteins include many redox proteins such as the berberine bridge enzyme‐like (BBE‐like) protein MoSef1. We further found that MoSef1 functions as an apoplastic virulent factor that inhibits the host immune response.

• Our studies revealed a novel function of a COPII uncoating factor in vesicular transport that is critical in the suppression of host immunity and pathogenicity of M. oryzae.

Introduction

Membrane‐coated vesicles and their transport are vital to the growth and development of many eukaryotic cells. Vesicle transport includes both endocytosis (uptake) and exocytosis (secretion) in which endocytosis starts from, and exocytosis ends at, the plasma membrane (PM) (Bonifacino & Glick, 2004; Shoji et al., 2014). The vesicle transport pathway comprises a series of complex and highly organised vesicular proteins that maintains the dynamic balance required for normal cellular growth and development. In plant pathogenic fungi, including the rice blast fungus Magnaporthe oryzae, vesicle transport is also critical in pathogenicity (Vida & Emr, 1995; Dou et al., 2011; Qi et al., 2016).

Clathrin‐mediated endocytosis (CME) is one of the endocytic processes in which foreign substances are taken up into clathrin‐coated pits through membrane invagination and transported to the endosomes in clathrin‐coated vesicles (CCVs) (McMahon & Boucrot, 2011). At the target compartment, CCVs become uncoated through vesicle disassembly and the cargo content is delivered through membrane fusion (McMahon & Boucrot, 2011; Weinberg & Drubin, 2012). In mammals, the disassembly of CCVs is regulated by the CME chaperone auxilin and heat shock protein Hsp70 (Fotin et al., 2004; Xing et al., 2010). Similarly, in the budding yeast Saccharomyces cerevisiae, this process is governed by the auxilin‐like Swa2 and heat shock protein Ssa1 (Xiao et al., 2006; Krantz et al., 2013). It was recently found that Swa2 coprecipitates with both COPI and COPII subunits in the early secretory pathway (Krantz et al., 2013).

Secretion starts from the endoplasmic reticulum (ER), the Golgi apparatus and then the PM to release cargo contents to the extracellular matrix. Various vesicular proteins were identified in the secretion process, with the most important ones being the coat protein complexes (COP) I and II (Barlowe & Miller, 2013). In comparison with COP I vesicles that facilitate ER to Golgi transport, COP II vesicles export the newly synthesised proteins from the ER (Barlowe & Miller, 2013). COP II is composed of five proteins, including Sar1, Sec23, Sec24, Sec13 and Sec31 that dimerise to form larger protein complexes (Gurkan et al., 2006).

Previous studies have indicated that M. oryzae secretes a large number of effector proteins into the host at the early stages of infection by two distinct secretion systems (Mosquera et al., 2009; Khang et al., 2010; Giraldo et al., 2013). Cytoplasmic effectors, such as Pwl2, Avr‐Piz‐t and Avr‐Pia, are delivered into the host via the biotrophic interfacial complex (BIC) mediated by t‐SNARE MoSso1 (Sweigard et al., 1995; Khang et al., 2008; Li et al., 2009; Chuma et al., 2011). Apoplastic effectors, such as Slp1 and Bas4 (Khang et al., 2010; Mentlak et al., 2012), are secreted into the membrane compartment of invasive hyphae (IH) (Giraldo et al., 2013). Despite this knowledge, how the effectors are transported to these sites is not clear. A previous study indicated that MoEnd3 is an endocytosis transport factor required for the secretion of Avr‐Pia and AvrPiz‐t, but not AvrPib and AvrPi9 (X. Li et al., 2017). Meanwhile, MoSyn8, a Qc‐SNARE protein, was also found to regulate cytoplasmic effectors’ secretion, but not the apoplastic effectors (Qi et al., 2016).

These above results collectively suggest that vesicular transport may play an important role in the secretion of various effector proteins, contributing to suppressing host immunity and virulence. In a previous study that characterised the importance of the actin‐regulating kinase MoArk1 in IH growth and host colonisation (Wang et al., 2013a), we identified a MoArk1‐interacting protein, MoSwa2, as an orthologue of yeast Swa2. We here show that MoSwa2 functions as an uncoating factor of COP II required for effector secretion and IH growth. We also show that MoSwa2 governs the secretion of MoSef1, a redox protein and apoplastic virulent factor, that inhibits host immunity.

Materials and Methods

Fungal strains and cultures

Magnaporthe oryzae strain Guy11 was used as the wild‐type strain in this study. The ∆Moark1 knockout mutant has been characterised and described previously by our laboratory (Wang et al., 2013a).

Assays for vegetative growth and conidiation

For vegetative growth, small agar blocks (2 × 2 mm) were cut from the edge of 4‐d‐old cultures and placed onto fresh medium (CM), oat medium (OM) or corn agar medium (SDC) for culture in the dark at 28°C for 7 d. For conidia production, mycelia were grown in the dark on SDC or CM medium at 28°C for 7 d, followed by constant illumination for 3 d. Conidia were harvested from 10‐d‐old cultures, filtered through three layers of lens paper and resuspended to a concentration of 5 × 10⁴ spores per ml in sterile water.

Yeast‐two‐hybrid assays

MoSWA2 full‐length cDNA was cloned into pGBKT7 as the bait construct. The prey construct was generated by cloning MoARK1 cDNA into pGADT7. The bait and prey constructs were confirmed by sequencing analysis and transformed into yeast strain AH109 using the BD library construction and screening kit (Clontech). Tryptophan (Trp⁺) and leucine (Leu⁺) transformants were isolated and assayed for growth on SD‐Trp‐Leu‐His (histone)‐Ade (adenine) medium and by expression of the LacZ reporter gene (Stratagene) (Yin et al., 2020).

In vivo co‐IP assays

The MoARK1‐3xFLAG plasmid containing the hygromycin‐resistance gene was constructed in our previous work (L. Li et al., 2017). The MoSWA2 DNA fragment fused with GFP fluorescent protein (MoSWA2‐GFP) was inserted into the pYF11 construct containing the bleomycin resistance gene. The constructs were co‐transformed into wild‐type strain Guy11 and transformants resistant to hygromycin and bleomycin were isolated. Total proteins were extracted from the transformants using protein lysis buffer (50 mM Tris‐HCl, pH 7.4, with 150 mM NaCl, 1 mM EDTA and 1% Triton X‐100) and incubated with Anti‐Flag^® M2 Affinity Gel (Sigma) for 4 h, followed by washing the Affinity Gel with Tris‐buffered saline (TBS) (50 mM Tris‐HCl, 150 mM NaCl, pH 7.4) four times. The proteins that bound to the Affinity Gel were eluted by 0.1 M glycine HCl (pH 3.5) and were detected by anti‐Flag (Sigma‐Aldrich) and anti‐GFP (Abmart) antibodies.

Phosphorylation analysis

For in vivo phosphorylation analysis, the MoSwa2‐GFP fusion constructs were transferred into the ΔMoswa2 and ΔMoark1 mutants, respectively. Positive transformants were cultured in liquid CM for 48 h. Next, c. 150–200 mg of mycelia were ground into a powder in liquid nitrogen and resuspended in 1 ml of extraction buffer (10 mM Tris‐HCl (pH 7.5), 150 mM NaCl, 0.5 mM EDTA, 0.5% NP40) to which 1 mM PMSF, 10 μl of protease inhibitor cocktail (Sigma) and 10 μl of phosphatase inhibitor cocktail 3 (Sigma) were freshly added. For phosphatase‐treated cell lysates, 2.5 U ml⁻¹ alkaline phosphatase (final concentration; P6774, Sigma) was added and the sample was incubated for 1 h with the addition of 1 mM MgCl₂ (37°C). Phosphatase inhibitor cocktail was not added. Samples were resolved on 8% sodium dodecyl sulfate (SDS)–polyacrylamide gels prepared with 50 μM acrylamide‐pendant Phos‐tag ligand and 100 μM MnCl₂ as described previously (L. Li et al., 2017). Gels were electrophoresed at 80 V/gel for 3–6 h. Before the transfer, gels were first equilibrated in transfer buffer containing 5 mM EDTA for 20 min two times and then in transfer buffer without EDTA for 10 min. Protein transfer from the Mn²⁺‐phos‐tag^TM acrylamide gel to the PVDF membrane was performed overnight at 80 V at 4°C and then the membrane was analysed by western blotting (using anti‐GFP antibodies).

For in vitro phosphorylation analysis, the GST‐MoArk1 and His‐MoSwa2 were expressed in E. coli strain BL21 (DE3) (Sigma, CMC0014) and purified as described in previous work (Liu et al., 2019). The reaction mixture was assembled by adding the following reagents to microcentrifuge tubes: purified recombinant kinase MoArk1 (0.5 μg) and potential substrate MoSwa2 (10 μg), 3 μl of 10× reaction buffer (250 mM Tris, pH 7.5, 100 mM MgCl₂ and 10 mM DTT) and water to 27.5 μl. The reaction was initiated by adding 2.5 μl of 0.5 mM ATP (Sigma‐Aldrich, FLAAS), mixed well and incubated at 22°C for 1.5 h, 10‐fold cold acetone was added to stop the reaction. Rapid and cost‐effective fluorescence detection in tube (FDIT) method was used to analyse protein phosphorylation in vitro (Yin et al., 2020).

Staining and confocal microscopy

For light microscopy studies, strains were grown on a thin layer of CM agar on microscope slides. After 2 d in a 28°C growth chamber, the hyphae were stained with N‐(3‐triethylammoniumpropyl)‐4‐(β‐diethylamino‐phenyl‐hexatrienyl)pyridinium dibromide (FM4‐64) (Molecular Probes Inc., Eugene, OR, USA) following procedures described previously (Qi et al., 2016). Briefly, FM4‐64 (1.3 mg ml⁻¹ in DMSO) was diluted in distilled H₂O at a final concentration of 5 μg ml⁻¹. Hyphae were wash with distilled water and stained with FM4‐64 on a glass slide in the dark. Hyphae were again washed with H₂O before observation. Confocal microscopy was performed using a Zeiss Axiovert 200M microscope equipped with a Zeiss LSM 710 META system. Excitation/emission wavelengths were 488 nm/505 nm for EGFP and 543 nm/560 nm for FM4‐64, respectively. Images were acquired and processed using LSM 710 AIM v.4.2 SP1 software (Zeiss, Oberkochen, Germany).

RNA isolation and quantitative RT‐PCR assay

Total RNA was isolated using the RNA simple Total RNA Kit (TIANGEN, China). All RNA was treated with DNase I (TaKaRa) before cDNA synthesis. First‐strand cDNA was synthesised using M‐MLV Reverse Transcriptase (Invitrogen) and oligo(dT) primers (TaKaRa).

Quantitative PCR was performed using the ABI 7500 real‐time PCR system in accordance with the manufacturer's instructions. The quantitative PCR reaction was carried out in a 20 μl volume containing 2 μl of reverse transcription product, 10 μl of SYBR^® Premix Ex Taq™ (2 μl), 0.4 μl ROX reference dye (50 μM) (SYBR^® PrimeScript™ RT‐PCR Kit, TaKaRa) and 0.4 μl of each primer (10 μM). To detect rice pathogenesis‐related (PR) gene transcription during the invasion growth stage, total RNA was extracted from plants inoculated with the wild‐type strain or mutant after 8, 24, 48 and 72 hpi. Transcription of the host ACTIN gene (XM_015774830.2) was used as the endogenous control (Supporting Information Table S1). The primer design was similar to that previously reported (Wang et al., 2017).

Target gene deletion and complementation

MoSWA2, MoSEC24‐2 and MoSEF1 gene deletion mutants were generated using the standard one‐step gene replacement strategy (Zhang et al., 2011). Two independent mutants with a similar phenotype were obtained and one of the randomly selected mutants was used for follow‐up characterisation. Fragments with 1.0 kb of sequences flanking the targeted genes were amplified by PCR with primer pairs. The resulting PCR products of target genes were digested with restriction endonucleases and inserted into the pCX62 vector containing a hygromycin‐resistance cassette (HPH) gene. The 3.4‐kb fragment containing the flanking sequences and the HPH cassette was amplified and introduced into Guy11 protoplasts. Putative transformants were verified by PCR and DNA sequencing and confirmed by Southern blotting analysis. The complement fragments, containing the native promoter region and the entire coding region, were amplified by PCR with primers and inserted into pYF11 (R25001; Thermo Fisher Scientific). Complement vectors were transformed into mutant protoplasts and screened by bleomycin resistance.

Virulence assays

For the virulence test, conidia were suspended to a concentration of 5 × 10⁴ spores ml⁻¹ in a 0.2% (w/v) gelatin solution and 5 ml of each were sprayed onto 2‐wk‐old rice seedlings (Oryza sativa CO39). For injection assay, conidia were suspended to a 1 × 10⁵ spores ml⁻¹ concentration and injected into the rice stem using a 1 ml syringe. Inoculated plants were kept in a growth chamber at 25°C with 90% humidity and in the dark for the first 24 h, followed by a 16 h : 8 h, light : dark cycle (Liu et al., 2019). Disease severity was assessed at 7 d after the inoculation.

Invasive growth and localisation assays

For observation of the penetration and invasive growth in rice cells, conidial suspensions (1 × 10⁵ spores ml⁻¹) of Guy11, mutants and complement strains were injected into rice leaf sheath. The inner epidermises of infected sheaths were harvested at different hours postinoculated by wild‐type and mutant strains and observed under a microscope. The complement strains with GFP fluorescent protein in the infected rice cells were observed using fluorescence microscopy (Zeiss Axio Observer A1 inverted microscope, ×63 magnification oil objective). Observation of effector secretion was performed as in our previous work (Qi et al., 2016). Briefly, for construction of Avr‐Pia:GFP, Avr‐Piz‐t:GFP and Bas4:GFP were transformed into Guy11, the ∆Moswa2 mutant and the ∆Mosec24‐2 mutant. The conidial suspensions were collected and followed the above infection assay. GFP fluorescence was observed using a fluorescence microscope (Zeiss Axio Observer A1 inverted microscope, ×63 magnification oil objective).

imagej analysis

Fluorescence was measured using imagej software. Briefly, Image‐Color‐Split Channels were extracted after opening the image, the RGB format was divided into 8‐bit format. The threshold was adjusted by choosing the Image‐Adjust‐Auto Threshold. Parameters were set to ensure that mean grey value and limit to threshold were checked.

Reactive oxygen species (ROS) observation

To observe ROS derived from the host, rice leaves or sheaths were stained with DAB (Sigma‐Aldrich), as described previously (Guo et al., 2011). Rice sheaths inoculated with the mutant and wild‐type strains for 24 hpi were incubated in 1 mg ml⁻¹ DAB solution, pH 3.8, at room temperature for 8 h and destained with clearing solution (ethanol : acetic acid, 94 : 4, v/v) for 1 h. The inner epidermises of infected sheaths were harvested and observed under a microscope.

ROS levels were monitored by luminol chemiluminescence assay (Liu et al., 2019). Rice protoplasts prepared, or leaves were cut into discs with a cork borer, and preincubated overnight in sterile‐distilled water in a 96‐well plate. The water was replaced with luminol (35.4 μg ml⁻¹) and peroxidase (10 μg ml⁻¹) solution and 50 nM purified mycelia (PRM) were used as the elicitor. Luminescence was measured using a GLOMAX96 microplate luminometer (Promega, Madison, WI, USA).

Organelle isolation

Protoplasts were prepared maintained in a buffer containing 100 mM Tris‐HCl, 0.1 mM MgCl₂, 10 mM DTT and 1.1 M sorbitol plus proteinases inhibitor mix (8215, Sigma‐Aldrich), as previously described (Guo et al., 2011). Organelles were isolated according to the established protocols with minor modifications as follows (Dunkley et al., 2004). Briefly, protoplasts were lysed in a buffer containing 10 mM Tris‐HCl, 0.5 mM MgCl₂, 8% Ficoll proteinases inhibitor mix and the lysate were placed on the top of a centrifuge tube containing a bottom layer of the lysis buffer (10 ml) and a top layer of the same buffer containing 4% Ficoll (10 ml). Intact organelles, which are viewed as an opaque layer at the top of the centrifuge tube following centrifugation (50 000 g, 45 min), were collected. The supernatants were loaded into a cushion of 18% iodixanol (Sigma‐Aldrich) for density gradient self‐generation. After precipitation at 100 000 g for 2 h in a swinging bucket rotor, the crude membranes were collected at the interface, adjusted to 16% in iodixanol and spun again at 350 000 g for 3 h. Fractions of 0.5 ml were harvested and washed with 0.8 ml 160 mM Na₂CO₃ for 30 min at 4°C. The membrane fraction was precipitated (100 000 g), washed with water and then precipitated again. Membranes were solubilised in a buffer (0.1 ml of 25 mM triethylammonium bicarbonate/8 M urea/2% Triton X‐100/0.1% SDS) and concentrations determined.

Preparation of vesicular proteins

Vesicle samples were prepared in accordance with an established work but with minor modifications (Wang et al., 2013b). About 5 g mycelia in liquid nitrogen were ground with a mortar and pestle and the mycelium powder was suspended in 25 ml of 0.1 M sodium acetate containing 0.07% β‐mercaptoethanol, then stirred at 4°C for 2 h. The mixture was centrifuged at 6500 g for 30 min and the supernatant was transferred to a new centrifuge tube. The precipitation was resuspended in 25 ml 0.1 M sodium acetate containing 0.07% β‐mercaptoethanol, after being distributed well, the mixture was centrifuged at 6500 g for 30 min, the supernatant was mixed with the previous preparation. The mixed supernatants were centrifuged at 7500 g for 30 min at 4°C to remove any debris. The supernatant was precipitated at 50 000 g for 90 min. Following ultracentrifugation, the supernatant was removed and dissolved in 1 ml of Tris‐MgCl₂‐DTT buffer (50 mM Tris pH 7.5, 10 mM MgCl₂, 5 mM DTT). After 30 min, the mixture was centrifuged at 8000 g for 3 min at 4°C and glycerol was added to the supernatant to a final concentration of 15%. For protein extraction, 400 μl vesicle samples were dissolved in 200 μl of lysis buffer (5 M urea, 2 M thiourea, 2% CHAPS, 2% SB 3–10, 40 mM Tris, 0.07% β‐mercaptoethanol, 1 mM PMSF). The proteins were precipitated by the addition of two volumes of pre‐cold acetone containing 10% TCA and 0.07% β‐mercaptoethanol, kept at −20°C for at least 30 min and centrifuged again at 13 000 g, 4°C for 30 min. The pellet was washed with cold acetone three times and air dried. Proteins were re‐dissolved in 200 µl of lysis buffer and centrifuged at 13 000 g to remove the insoluble impurity.

Transmission electron microscopy observation

For transmission electron microscopy (TEM) observation, vegetative hyphae of indicated strains were cultured in liquid complete medium (CM) for 32 h. The hyphae were then fixed with 2.5% glutaraldehyde in phosphate buffer (pH 7.0) for more than 4 h, washed three times in phosphate buffer and fixed with 1% OsO₄ in phosphate buffer (pH 7.0) for 1 h followed by three washes with the phosphate buffer. The specimen was dehydrated by a gradual series of ethanol (30%, 50%, 70%, 80%, 90% and 95%) for c. 15 min at each step and then transferred to and incubated in absolute acetone for 20 min. The sample was then placed in the 1 : 1 mixture of absolute acetone and the Spurr resin mixture for 1 h at room temperature. The mixture was again transferred to a 1 : 3 mixture of absolute acetone and the resin mixture for 3 h and to finalise Spurr resin mixture overnight. The sample was placed in capsules containing embedding medium and heated at 70°C for 9 h. The sample was sectioned in LEICA EM UC7 Ultratome and sections were stained with uranyl acetate and alkaline lead citrate for 5–10 min, respectively. Observation was made using a Hitachi Model H‐7650 TEM.

Results

MoSwa2 is subjected to phosphorylation regulation by MoArk1 and is required for normal endocytosis

In a previous study that characterised the importance of the actin‐regulating kinase MoArk1 in endocytosis, invasive growth and pathogenicity of M. oryzae, we have identified MoSwa2 (MGG_04080) as an orthologue of Swa2 in S. cerevisiae that interacts with MoArk1 (L. Li et al., 2017). As the yeast Swa2 is an auxilin‐like protein and a clathrin disassembly factor required for membrane transport in endocytosis pathway (Xiao et al., 2006; Xing et al., 2010), we thought to characterise the MoSwa2 function in membrane transport and virulence of the blast fungus.

We first confirmed the interaction between MoSwa2 and MoArk1 by yeast‐two‐hybrid (Y2H) and co‐immunoprecipitation (co‐IP) assays (Fig. 1a,b). We then found that MoSwa2 shared a high similarity with other fungal Swa2 proteins, such as Gaeumannomyces graminis XP_009224220.1, Neurospora tetrasperma XP_009847368.1 and Neurospora crassa XP_956212.1. Moreover, MoSwa2 contains a UBA domain, a conserved Ark/Prk kinase phosphorylation target motif [L/I/V]XXXXTG‐containing TPR domain (Smythe & Ayscough, 2003; Wang et al., 2009) and a C‐terminus DnaJ domain, similar to Swa2 of other fungi (Fig. S1). To test whether MoSwa2 was regulated by MoArk1 through protein phosphorylation, Mn²⁺‐Phos‐tag SDS‐PAGE was performed. We extracted the MoSwa2:GFP protein from the wild‐type Guy11 strain expressing the MoSWA2:GFP fusion construct, treated with either a phosphatase or a phosphatase inhibitor and performed Mn²⁺‐Phos‐tag SDS‐PAGE. The band for MoSwa2:GFP treated with the inhibitor migrated more slowly than that treated with phosphatase, indicating the occurrence of in vivo phosphorylation of MoSwa2 in Guy11. By contrast, a similar migration was found between MoSwa2:GFP of ΔMoark1/MoSWA2:GFP and of Guy11/MoSWA2:GFP (Fig. S2). We inferred that MoSwa2 may be subject to additional phosphorylation mechanisms. To test this possibility, we utilised the FDIT method (Jin & Gou, 2016; Yin et al., 2019). GST‐MoArk1, His‐MoSwa2 and phosphorylation site inactivation mutation His‐MoSwa2^T670A proteins were expressed in the Escherichia coli strain BL21. A significant increase in fluorescence was detected in GST‐MoArk1/His‐MoSwa2/ATP, but not in the control (Fig. 1d), suggesting that MoArk1 can phosphorylate MoSwa2.

Fig. 1.

MoSwa2 is subjected to phosphorylation regulation by MoArk1 and is required for normal endocytosis. (a) Yeast‐two‐hybrid assay for the interaction between MoSwa2 and MoArk1. Yeast transformants expressing the prey and bait constructs were assayed for growth on SD−Leu−Trp and SD−Leu−Trp−His plates with β‐galactosidase activities (LacZ). (b) Co‐immunoprecipitation (co‐IP) analyses of the interaction between MoSwa2 and MoArk1. The MoArk1‐3xFlag and MoSwa2‐GFP were co‐expressed in the wild‐type strain Guy11. The co‐IP experiment was performed with the anti‐Flag Affinity Gel and the isolated protein was analysed by western blotting using anti‐FLAG and anti‐GFP antibodies. (c) Domain map of MoSwa2. The protein sequence of MoSwa2 was used to perform the SMART (http://smart.embl‐heidelberg.de) and Motif Scan (https://myhits.sib.swiss/cgi‐bin/motif_scan) analysis service for functional domain prediction. (d) In vitro phosphorylation analysis by FDIT method. Purified proteins of GST‐MoArk1 and His‐MoSwa2 were used for protein kinase reaction and then dyed with Pro‐Q^® Diamond Phosphorylation Gel Stain. The fluorescence signal at 590 nm (excited at 530 nm) was measured in a Cytation3 microplate reader (Biotek, Winooski, VT, USA). The experiments were repeated three times and showed similar results. Asterisks indicate statistical significances (P < 0.01). (e) Time course images of FM4‐64 uptake at the hyphal tips. Hyphae stained by FM4‐64 were examined using fluorescence microscopy at different time points. Insets highlight areas analysed by line scan. Fluorescence intensity was measured using imagej software.

Gene deletion analysis revealed that MoSwa2 is required for normal endocytosis but dispensable for vegetative growth and conidiation (Table S2; Fig. S3). Weak fluorescence was observed in the PM of the ∆Moswa2 mutant after 5 min incubation with dye FM4‐64, in contrast with a strong signal in the endomembrane, consisting of vacuoles and endosomes, from the Guy11 and ∆Moswa2/MoSWA2 complemented strain. The constitutively phosphorylated MoSwa2^T670D strain, but not the dephosphorylated MoSwa2^T670A strain, partially restored the defect in FM4‐64 uptake (Fig. 1e). However, neither MoSwa2^T670A nor MoSwa2^T670D could rescue the virulence defect of the ∆Moark1 mutant (Fig. S4).

MoSwa2 is important for normal virulence and invasive hyphae growth

To test the role of MoSwa2 in virulence, the conidial suspensions of Guy11, the ∆Moswa2 mutant and the complemented strain (∆Moswa2/MoSWA2) were sprayed onto 2‐wk‐old susceptible rice cultivar CO39 at a concentration of 5 × 10⁴ conidia ml⁻¹. The ∆Moswa2 mutant showed significantly reduced lesions compared with the two other strains (Fig. 2a,b). We also estimated conidia formation on surface‐sterilised rice leaves (Qi et al., 2016). Leaf lesions (>80%, n = 100) by Guy11 and the complemented strain (>80%, n = 100) produced abundant conidia, but less than 20% was seen on the leaves inoculated with the ∆Moswa2 mutant (n = 73) (Fig. 2c). Based on whether conidiation was possible or not on the lesions (Park et al., 2002), we divided them into two categories, typical lesions and necrotic lesions (Fig. 2c). Abundant and typical lesions were found on leaves inoculated with Guy11 and the complemented strain at 7‐d postinoculation (dpi), but not those infected with the ∆Moswa2 mutant, which showed a few small necrotic lesions (Fig. 2c), indicating MoSwa2 contributed to the virulence of M. oryzae.

Fig. 2.

MoSwa2 is important for normal virulence. (a) Conidial suspensions of Guy11, ∆Moswa2 mutant and the complement strains were sprayed onto 2‐wk‐old rice seedlings (CO39). Diseased rice leaves were photographed after 7 d postinoculation (dpi). Typical lesions are evident upon inoculation with Guy11 and the complemented strain, whereas the ΔMoswa2 mutant shows necrotic lesions (inset). (b) Lesion numbers of the leaves infected by Guy11, ∆Moswa2 mutant and the complement strains were counted. Experiments were repeated three times and showed similar results. Asterisks indicate a significant difference (P < 0.01). Bars represent medians and boxes the 25^th and 75^th percentiles. (c) Statistics for conidiation lesions on surface‐sterilised rice leaves above. Error bars represent the SD.

To understand how the ∆Moswa2 mutant attenuated virulence, we performed an infection assay on rice sheath and observed IH growth at 100 appressorium penetration sites. We rated the hyphal growth as from level I to level IV at 30 h postinoculation (hpi) (I, no penetration; II, with primary invasion hyphal; III, secondary invasive hypha does not extend to neighbouring plant cells; IV, invasion hyphal extended into neighbouring plant cells) (Fig. 3a) (Qi et al., 2016). In the wild‐type and the complemented strains, 98% and 94% of penetration sites showed level III and level IV invasive growth. By contrast, 55% of penetration sites showed levels I and II and only 45% showed levels III and IV in the ∆Moswa2 mutant (Fig. 3b). Strikingly, the majority of the ∆Moswa2 mutant infected sites (>85%, n = 100) exhibited a highly specific and strong defence response. There were lots of dark and brown inclusions attached to the IH of the ∆Moswa2 mutant at 36 hpi, compared with those by the wild‐type strain (<10%, n = 100) (Fig. 3c,d). Moreover, expression of the PR genes PR1a and PBZ1 was highly induced at 48 and 72 hpi after ∆Moswa2 infection, compared with infection by the wild‐type strain (Fig. S5). We therefore concluded that ∆Moswa2 infection elicits a strong defence response from rice that restricts the growth of IH.

Fig. 3.

MoSwa2 is required for normal invasion growth. (a, b) Close observation and statistics for invasive growth in rice leaf sheath. Statistics of invasive hyphal growth at 100 appressorium penetration sites by rating the hyphal growth from level I to level IV (I, no penetration; II, with primary invasive hypha; III, secondary invasive hypha does not extend to the neighbouring plant cells; IV, invasive hypha extended into neighbouring plant cells). Experiments were repeated three times and showed similar results. (c, d) MoSwa2 has a role in suppressing host immunity. Highly localised defence response in rice marked with black inclusions and restricted growth of the ΔMoswa2 invasive hyphae at 36 h postinoculation (hpi). The percentage of the pattern shown in the image was calculated by observation for 100 penetration sites that were randomly chosen. Experiments were repeated three times and showed similar results. Error bars represent the SD from three independent replicates and asterisks indicate a significant difference (P < 0.01). Bars, 10 μm.

MoSwa2 is required for host ROS suppression

During plant–pathogen interactions, the pathogen often encounters plant defence barriers, such as the ROS (Guo et al., 2010; Caillaud et al., 2014). To test if the ∆Moswa2 mutant instigated such a response from the host, we estimated ROS production using DAB staining in infected rice sheaths at 24 h. Although ROS was rarely detected by Guy11 infection (<10%, n = 100), a strong presence was observed in cells penetrated by the ∆Moswa2 mutant (>80%, n = 100) (Fig. 4a,b).

Fig. 4.

MoSwa2 is important for suppressing the production of reactive oxygen species (ROS) in rice cells. (a, b) DAB staining on infected leaf sheath of CO39 with Guy11, ∆Moswa2 mutant and the complement strains at 24 h postinoculation (hpi). Graphics represent the number of appressoria (n = 100) that induce H₂O₂ in the invaded rice cells. Experiments were repeated three times and showed similar results. (c, d) Invasive growth of ∆Moswa2 mutant was restored by DPI and CAG treatment; 100 penetration sites of each sample were counted. Experiments were repeated three times and showed similar results. Error bars represent the SD and asterisks indicate a significant difference (P < 0.01). Bars, 10 μm.

To determine whether the strong ROS production in ∆Moswa2 infection was due to a lack of ROS scavengers in host cells, we tested the effect of pretreatment with DPI (diphenyleneiodonium, an inhibitor of NADPH oxidases) (Cross & Jones, 1986; Chen et al., 2014) and CAG (catalase of Aspergillus niger, a scavenger of H₂O₂) (Tanabe et al., 2009; Liu et al., 2016) by rating the IH growth of 100 appressorial penetration sites. In infections by the wild‐type and the complemented strains, c. 80% of penetration sites showed level II and level III IH growth with or without DPI and CAG at 24 h, respectively. By contrast, 80% of penetration sites showed level I and 15% level III and level IV IH growth in ∆Moswa2 infection. When treated with DPI or CAG, the restricted IH growth could be rescued at 24 h and 48 h (Fig. 4c,d). These results suggested that ROS bursts in rice cells infected by ∆Moswa2 mutant led to restricted invasive growth and virulence of M. oryzae.

MoSwa2 is involved in effector protein secretion

Plant pathogens often secrete effectors to interfere with host immunity during infection (Jones & Dangl, 2006; Zhang et al., 2016; Wang & Wang, 2018). To determine whether MoSwa2 is also involved in the secretion process, we first observed the subcellular localisation of MoSwa2 during various development stages. We found GFP fluorescence of MoSwa2 in vesicle‐like structures throughout the cytoplasm of normal hyphae, conidia and germ tubes (Fig. S6A). We further performed time‐lapse imaging to observe the localisation of MoSwa2 and found vesicle‐like structures with irregular movement in IH (Fig. S6B). As vesicle trafficking plays a crucial role in effector secretion (Qi et al., 2016; X. Li et al., 2017), these findings indicated that MoSwa2 may have a role in secretion.

To test if MoSwa2 affected the secretion of the effector proteins, AVR‐Pia and AVR‐Piz‐t genes fused with a C‐terminal GFP were expressed in Guy11, the ∆Moswa2 mutant and the ∆Moswa2/MoSWA2 complement strain. We found that the ∆Moswa2 mutant expressing Avr‐Pia or Avr‐Piz‐t produced typical lesions on LTH‐Pia (Pia R gene monogenetic rice line) or LTH‐Piz‐t (Piz‐t R gene monogenetic rice line) rice lines, in contrast with Guy11 and the complement strain that did not produce any lesions (Fig. 5a,b,d,e). This result indicated that MoSwa2 interferes with the recognition of Pia/Avr‐Pia and Piz‐t/Avr‐Piz‐t.

Fig. 5.

MoSwa2 is involved in the secretion of effector proteins. (a, d) Whole‐plant assays with cultivars LTH, LTH‐Pia or LTH‐Piz‐t inoculated with Guy11^a (Guy11 expressed Avr‐Pia), ΔMoswa2^a (ΔMoswa2 expressed Avr‐Pia), Guy11^b (Guy11 expressed Avr‐Piz‐t) or ΔMoswa2^b (ΔMoswa2 expressed Avr‐Piz‐t) and the complement strains, respectively. Incompatible Guy11^a, Guy11^b and the complement strains generally failed to form visible symptoms on LTH‐Pia and LTH‐Piz‐t, respectively, while they generated typical expanding lesions on the compatible cultivar LTH. The ΔMoswa2 ^a and ΔMoswa2^b mutant formed rare, typical lesions on LTH‐Pia and LTH‐Piz‐t, respectively, similar to those it produced on the susceptible cultivar LTH. Photographs were taken 7 d postinoculation (dpi). (b, e) Lesion numbers of the infected leaves shown in (a) and (d) were counted. Experiments were repeated three times and showed similar results. Asterisks indicate a significant difference (P < 0.01). Horizontal lines represent medians, ranges represent the 25^th and 75^th percentiles and the circles mean the samples. ‘nd’ means not detected. (c, f) In susceptible CO39 sheath cells, Avr‐Pia:GFP or Avr‐Piz‐t:GFP fluorescence was secreted and accumulated in BIC (arrowhead) in Guy11 but not in ΔMoswa2 mutant. Photographs were taken at 24 h postinoculation (hpi) with Guy11 and the complement strains and 32 hpi with ΔMoswa2 mutant. Bars, 10 μm.

We further observed the localisation of Avr‐Pia and Avr‐Piz‐t in Guy11 and the ∆Moswa2 mutant during infection. GFP was detected in the BICs of Guy11 (>80% of 100 imaged infection sites), but not the BICs of the ∆Moswa2 mutant (Fig. 5c,f). Moreover, examination for expression showed that disruption of MoSWA2 did not significantly affect Avr‐Pia and Avr‐Piz‐t transcription (Fig. S7). We then detected the localisation of the apoplastic effector Slp1. GFP was readily detected in the extrainvasive hyphal membrane (EIHM) of Guy11 (>80% imaged of 100 infection sites), but not the ∆Moswa2 mutant in which GFP appeared to be intermittent and diffused in the cytoplasm (Fig. S8).

We further tested whether MoSwa2 phosphorylation by MoArk1 affected effector secretion and found that neither the constitutively phosphorylated MoSwa2^T670D strain nor the dephosphorylated MoSwa2^T670A strain could rescue the defects of the ∆Moswa2 mutant in secretion of effectors and virulence (Figs S9,S10). We also found that MoArk1 is dispensable for secretion of effectors (Fig. S9). Taken together, the above findings indicated that MoSwa2 was required for secretion of the cytoplasmic effectors Avr‐Pia and Avr‐Piz‐t and the apoplastic effector Slp1. Moreover, this role appears to be independent of its phosphorylation by MoArk1.

MoSwa2 functions as an uncoating factor required for COP II‐mediated secretion

In S. cerevisiae, Swa2 functions as the uncoating factor for CCVs and it also links to COP II and COP I vesicles in the early secretory pathway (Ding et al., 2016). To examine if MoSwa2 exhibited the uncoating function for COP vesicles that underlies its role in effector secretion of M. oryzae, we first initiated binding studies with the COP II inner shell complex proteins MoSec23, MoSec24‐1 and MoSec24‐2, the outer shell complex proteins MoSec13 and MoSec31, as well as the COP I cage subunits MoSec14‐1, MoSec14‐2, MoSec26 and MoSec28. We constitutively expressed the COP II and COP I subunits fused with a C‐terminus Flag‐tag in the complemented strain (∆Moswa2/MoSWA2‐GFP) and found that only the Sec24‐2‐Flag protein showed a strong interaction with the anti‐GFP antibody (Fig. 6a). This result indicated that MoSwa2 might interact with the COP II coat complex.

Fig. 6.

MoSwa2 functions as an uncoating factor required for COPII vesicles. (a) MoSwa2 interacts with COPII subunits in vivo. Co‐IP analyses of the interaction between MoSwa2 and COP subunits. The COP subunits fused with Flag‐tag and MoSwa2‐GFP were co‐expressed in the wild‐type strain Guy11. The co‐IP experiment was performed with the anti‐GFP agarose and the isolated protein was analysed by western blotting using anti‐FLAG and anti‐GFP antibodies. (b) Organelles from M. oryzae protoplasts were partially separated by centrifugation through self‐generating Ficoll density gradients, as shown in the diagram. The ER and Golgi apparatus distribution and gradient fractions were analysed by western blotting, using RFP antibodies against the Golgi marker MoSft2 and the ER marker MoLhs1 fused with RFP. Distribution of MoSec24‐2‐Flag was detected by anti‐Flag antibody. The red dotted box indicates abundant fractions. (c) A comparison of the vesicle proteins (V) and other cytoplasmic components (C) of Guy11 and ∆Moswa2 mutant expressing MoSec24‐2‐Flag as displayed by western blotting. Relative band intensity was quantified by imagej software. Protein loading is indicated with Coomassie brilliant blue (CBB) stain. (d) Transmission electron microscopy observes the hyphae of Guy11 and ∆Moswa2 mutant. Electron‐transparent vesicles are marked by arrows. Bars, 1 µm, or 0.2 µm.

Moreover, we obtained a ∆Mosec24‐2 knockout mutant strain to test the hypothesis that a blocked uncoating process may lead to a disordered distribution of the COP II subunits in the ER–Golgi. We expressed MoSec24‐2‐Flag in Guy11 and the ∆Moswa2 mutant and fractionated subcellular organelles by gradient centrifugation. Each fraction was analysed by western blotting using the red fluorescent protein (RFP) antibody against RFP‐tagged MoSft2 (a Golgi marker) and MoLhs1 (an ER marker) (Yi et al., 2009a; Zhang et al., 2019). The Golgi were more abundant in fractions 4, 5 and 6 and the ER was more abundant in fractions 3 and 4 (Fig. 6b). We found that MoSec24‐2 was mainly distributed in the Golgi and ER fractions of Guy11, but its distribution was different in the ∆Moswa2 mutant (Fig. 6b). This result indicated that MoSec24‐2 might be retained in the COP II complex due to a defective uncoating process. To further explore this finding, we compared the subcellular distributions of MoSec24‐2‐Flag between Guy11 and the ∆Moswa2 mutant. We found that MoSec24‐2 was mainly concentrated in the vesicles of the ∆Moswa2 mutant, but not the cytoplasm, as seen in Guy11 (Fig. 6c). This observation was further corroborated by TEM that revealed more vesicles produced by the ∆Moswa2 mutant than by Guy11 (Fig. 6d). All this evidence suggested that MoSwa2 affected the disassembly of COP II by interacting with MoSec24‐2.

We subsequently performed live‐cell imaging of hyphae, appressoria and IH from Guy11 and the ∆Moswa2 strains expressing the MoSec24‐2‐GFP fusion protein. MoSec24‐2 fluorescence was mostly concentrated in the periphery of hyphal tips in Guy11, but dispersed in the ∆Moswa2 mutant cytoplasm. In the appressorium, fluorescence was detected in a toroidal‐shaped structure in Guy11, but not the ∆Moswa2 mutant. Consistently, the location of MoSec24‐2 in IH was not evenly distributed in the cytoplasm of Guy11, but several aggregations in the cytoplasm of the ∆Moswa2 mutant (Fig. S11). Finally, we found that MoSec24‐2 played a similar role as MoSwa2 in the secretion of effector proteins Avr‐Pia, Avr‐Piz‐t and Slp1 (Fig. S12), virulence (Fig. S13a–c) and IH growth (Fig. S13d). Collectively, these results demonstrated that MoSwa2 functioned as an uncoating factor of COP II and together with MoSec24‐2, MoSwa2 facilitated the secretion of effectors to attenuate virulence in M. oryzae.

MoSwa2 regulates extracellular protein secretion

M. oryzae secretes various extracellular proteins during the early stages of infection (Y. Wang et al., 2011; Liu et al., 2018). To test if MoSwa2 had a role in protein secretion, we extracted extracellular fluid (EF) from Guy11 (Patkar et al., 2015; Liu et al., 2018) and supplemented this in the rice leaf sheath inoculated with the ∆Moswa2 mutant. A positive virulence reaction was observed (Fig. 7). When the EF of Guy11 was inactivated by boiling, a similar effect was not found (Fig. 7). To further identify the active EF components, we subjected the EFs of Guy11 and the ∆Moswa2 mutant to mass spectrometry analysis. Here, c. 122 secreted proteins were identified from Guy11 (Table S3) that could be functionally classified into five major categories: cell wall modification (31%), ROS modification (19%), protein modification (24%), RNA editing (2%) and unknown function (24%) (Fig. S14a). These proteins were not found in the EF of ∆Moswa2, suggesting that MoSwa2 governs the secretion of multiple proteins likely to be involved in the initial steps of host invasion.

Fig. 7.

MoSwa2 regulates the secretion of extracellular proteins. (a) Loss of MoSWA2 in M. oryzae leads to a strong immune response with inclusion deposition in rice. Extracellular fluid (EF) from Guy11 appressoria suppresses the ΔMoswa2 mutant phenotype, whereas the boiled (b) EF of Guy11 cannot rescue the defects in the ΔMoswa2 mutant. (b) Graphical representation of the number of appressoria that induce the characteristic defence response (inclusions), as shown in (a), in the invaded rice cells. Data represent observations from three independent experiments (n = 100 appressoria each). Error bars represent the SD. Bar, 10 μm.

MoSwa2 regulates MoSef1, an EF protein important for ROS detoxification

Fungal extracellular ROS‐detoxifying enzymes are associated with virulence during the pathogen–host interaction (Tanaka et al., 2006; Qin et al., 2007). To further test if MoSwa2 has a role in ROS detoxification, we examined the above differentially expressed 122 EF proteins for those with a role in ROS modification and successfully obtained 11 gene deletion mutants (Fig. S14b). To evaluate their roles in ROS production, conidial suspensions of each mutant were inoculated on rice leaf sheaths and DAB staining was carried out. Rice inoculated with ∆Mosef1 (MGG_09717, a MoSwa2 regulated EF protein 1) exhibited a higher level of H₂O₂ while four other mutant strains had moderately high accumulation of H₂O₂ and six other mutants were normal for H₂O₂ levels (Figs 8f,S15). Despite that it was dispensable for vegetative growth and conidiation (Table S2), the ∆Mosef1 mutant was attenuated in virulence (Fig. 8a–e). As an additional reference, virulence of the other 10 mutants were also tested and are shown in Fig. S15.

Fig. 8.

Extracellular protein MoSef1 is required for normal virulence and invasive growth. (a, c) For spray assay, conidial suspensions of Guy11, ∆Mosef1 mutant and the complement strains were suspended to a concentration of 5 × 10⁴ spores ml⁻¹ and sprayed onto 2‐wk‐old rice seedlings (CO39). For injection assay, conidia were suspended to a concentration of 1 × 10⁵ spores ml⁻¹ and injected into rice stems using a 1 ml syringe. Diseased rice leaves were photographed after 7 d postinoculation (dpi). (b) Lesion numbers of the infected leaves shown in (a) were counted. Experiments were repeated three times and showed similar results. Bars represent medians and boxes the 25^th and 75^th percentiles. Statistics for conidiation lesions on surface‐sterilised rice leaves above. Error bars represent the SD and the asterisks indicate a significant difference (P < 0.01). (d) Disease lesion areas were assessed by imagej software. There were three replicates for each of 10 leaves. Bars represent medians and boxes indicate the 25^th and 75^th percentiles. Fungal growth was evaluated by quantifying Magnaporthe oryzae genomic 28S rDNA relative to rice genomic Rubq1 DNA. The mean values of three determinations with SDs are shown. Asterisks indicate a significant difference (P < 0.01). (e) For observation of the penetration and invasive growth in rice cells, conidial suspensions (1 × 10⁵ spores ml⁻¹) were injected into rice leaf sheath. At 28°C for 24 or 48 h, the infected sheaths’ inner epidermises were observed under a microscope. The percentage of the pattern shown in the image was calculated by observation of 100 penetration sites that were chosen at random and observation was conducted three times. Bar, 10 μm. (f) DAB staining on infected leaf sheath of CO39 at 24 h postinoculation (hpi) with Guy11, ∆Mosef1 mutant and the complement strains. Bar, 10 μm. (g) Domain map of MoSef1. The protein sequence of MoSwa2 was used to perform the SMART analysis service. (h) MoSef1 inhibits reactive oxygen species (ROS) bursts induced by PRM in the CO39 rice plant. Rice protoplasts from CO39 were treated with purified MoSef1 protein, 100 nM PRM or water and ROS were detected with a luminol‐chemiluminescent assay. Error bars represent the SD (n = 3).

Sequence analysis revealed that MoSef1 contained an N‐terminus signal peptide, three chitin binding domains, a FAD binding domain and a C‐terminus berberine bridge enzyme (BBE) domain (Fig. 8g). Using a luminol‐based chemiluminescence assay (Liu et al., 2019), we observed that the in vitro expression of MoSef1 can reduce the PRM‐triggered increase of ROS (Fig. 8h). We also found that MoSef1 is an apoplastic effector localised in EIHM and the location requires the function of both MoSwa2 and MoSec24‐2 (Fig. S16). Collectively, these results indicated that MoSwa2 regulated EF proteins including MoSef1 that modulated the production of ROS detoxification enzymes.

Discussion

Cellular membranes can generate vesicles that carry cargo travel between various cellular compartments necessary for growth and development. Most trafficking vesicles are initially coated, including clathrin‐coated and COP I and COP II vesicles (Mallabiabarrena & Malhotra, 1995). It was reported that there are multiple coat adaptors in vesicle formation, transport and targeting (Cai et al., 2007; Angers & Merz, 2009; Ishikawa et al., 2016; Yuan et al., 2018). The J‐domain protein auxilin is a cofactor for Hsc70 in uncoating CCVs and it plays a role during CME (Lemmon, 2001; Eisenberg & Greene, 2007). The present study’s focus, the auxilin‐like protein MoSwa2, is known as an orthologue of Swa2, a clathrin uncoating factor and a phosphorylated protein of yeast (Gall et al., 2000). However, the kinase that phosphorylates Swa2 has not been previously identified. We found here that MoSwa2 is subjected to phosphorylation regulation by MoArk1 in endocytosis.

Previous studies revealed that there are two distinctive effector secretion systems in M. oryzae (Giraldo et al., 2013). The cytoplasmic effectors secrete to BIC involving exocyst components and the MoSso1 SNARE protein. The apoplastic effectors secreted into the EIHM follow a conventional secretory pathway (Khang et al., 2010; Giraldo et al., 2013). Exocyst components and SNARE proteins were identified as the docking submits in tethering vesicles to the secretory sites (Heider & Munson, 2012). In addition, the MoSyn8 Qc‐SNARE protein was involved in the secretion of cytoplasmic effectors, but not apoplastic effectors (Qi et al., 2016) and MoEnd3 was required for the secretion of cytoplasmic effectors Avr‐Pia and AvrPiz‐t, but not AvrPib and AvrPi9 (X. Li et al., 2017). All these findings indicated that the secretion of different effectors might depend on distinct vesicle transport pathways. Despite this knowledge, vesicle types, how they transport and their effectors remain unknown. Here, we found that MoSwa2 interacted with COP II subunit MoSec24‐2. MoSwa2 governed the disassembling of the COP II coats and disruption of the MoSWA2 gene results in mis‐localisation of MoSec24‐2. We also found that MoSwa2 functioned as an uncoating factor via interacting with the COP II subunit MoSec24‐2 to mediate the secretion of both cytoplasmic and apoplastic extracellular proteins that govern the pathogenicity of M. oryzae (Fig. 9).

Fig. 9.

Working model of the MoSwa2‐mediated secretion pathway. During clathrin‐mediated endocytosis (CME), clathrin‐coated vesicles (CCV) mainly function in extracellular material uptake from the surface into intracellular compartments, endosomes or vacuole (V) for degradation (McMahon & Boucrot, 2011). The uncoating process of CCV is regulated by auxilin‐like protein Swa2, allowing the detached and uncoated vesicle to move to and fuse with its target endosome (Fotin et al., 2004; Xing et al., 2010). During the early secretory pathway, the COPII‐associated process mainly functions in cargo transport from the endoplasmic reticulum (ER) to the Golgi apparatus (G). The COPII coat consists of an inner shell (Sec23/Sec24) that sorts cargo and an outer shell (Sec13/Sec31) (Gurkan et al., 2006). There are two distinct secretory pathways for effector proteins secretion in Magnaporthe oryzae, Cytoplasmic effectors are delivered into the host via BIC and apoplastic effectors are secreted into the EIHM. Both pathways rely on vesicle transport initiated from ER (Khang et al., 2010; Giraldo et al., 2013). Here, MoSwa2 functions as an uncoating factor for CCV coats via phosphorylation by actin‐regulating kinase MoArk1, which displays conserved functions important in endocytosis via substrate phosphorylation in thee CME process (X. Li et al., 2017). MoSwa2 also functions as an uncoating factor for COPII in the early secretory pathway, via interacting with COPII subunits MoSec24‐2 to mediate cytoplasmic and apoplastic extracellular effector secretion to inhibit the host immune response.

It is well known that ROS production in the early infection stage induces the inclusion accumulation that stimulates the expression of PR genes (Tanaka et al., 2003; Torres et al., 2005; Chi et al., 2009; Yi et al., 2009b). Note that the plant defence responses against ∆Moswa2 were regulated by the ROS level, because not only did treatment with ROS inhibitors DPI and CAG result in a dramatic reduction of the plant defence response on rice, but also expression of PR1a and PBZ1 was induced in the mutant. For successful colonisation, pathogens have developed ROS scavenging mechanisms, such as the secretion of extracellular redox enzymes, to suppress the host defence response. Studies of secreted extracellular proteins have emphasised their importance in mediating microbe–host interactions (Q. Wang et al., 2011; Dong et al., 2015). A previous study found that an amino‐phospholipid translocase 2 mediated many enzymes’ secretion and induced a host defence response in an incompatible interaction (Gilbert et al., 2006). Additional studies found that Hsp70 family proteins, KAR2 and LHS1, were required for the secretion of effectors and virulence in M. oryzae (Yi et al., 2009a). The Rab GTPase MoSec4 also plays an important role in the secretion of extracellular proteins for pathogenicity (Zheng et al., 2016). The Ustilago maydis effector protein Pep1 inhibited the host extracellular enzyme POX12 to suppress ROS production (Hemetsberger et al., 2012). Here, we found that MoSwa2 was involved in the secretion of Avr‐Pia, Avr‐Piz‐t and Slp1 that are involved in targeting the host ubiquitin–proteasome system (Park et al., 2012, 2016; Wang et al., 2016) or suppressing chitin‐triggered host immunity (Mentlak et al., 2012). MoSwa2 not only regulated the secretion of Avr proteins but also a repertoire of extracellular proteins. Among 122 differentially expressed EF proteins, we found that 19% were extracellular enzymes participating in redox modification. Specifically, we found that a putative apoplastic BBE‐like factor, MoSef1, was important in fungus virulence. Loss of MoSEF1 induced a strong accumulation of ROS in infected cells and in vitro produced MoSef1 suppressed ROS in rice leaves. In Arabidopsis thaliana, BBE‐like enzymes function as oligogalacturonide (OG) oxidases to reduce the capability of OGs in activating the host immune response (Benedetti et al., 2018). In Phytophthora infestans, a predicted secretome revealed that BBE‐like enzymes were required for host invasion (Raffaele et al., 2010). Interestingly, despite an apparent role in modulating ROS response and virulence, MoSef1 is dispensable for vegetative growth and conidiation. Further studies of the MoSef1 infection‐specific function, including how it interferes with host immunity, are highly warranted.

Author contributions

ML and ZZ designed research; ML, JH, YH, AZ, YD, WC and ZZ performed experiments; ML and ZZ contributed new reagents/analytical tools; ML, JH, YH, HZ and XZ analysed data; and ML and ZZ wrote the manuscript.

Supporting information

Fig. S1 Phylogenetic analysis of putative Swa2 proteins in fungi.

Fig. S2 MoSwa2 is phosphorylated in vivo.

Fig. S3 Southern hybridisation.

Fig. S4 Constitutively phosphorylated or dephosphorylated MoSwa2 cannot rescue ΔMoark1 defects.

Fig. S5 Expression of plant PR genes in response to infection by the wild‐type and the ΔMoswa2 mutant.

Fig. S6 Subcellular localisation of MoSwa2 in different development stages.

Fig. S7 Expression of AVR‐Pia and AVR‐Piz‐t of wild‐type and the deletion mutant in the infection stage.

Fig. S8 MoSwa2 is involved in the secretion of Slp1.

Fig. S9 Phosphorylation of MoSwa2 by MoArk1 is dispensable for the secretion of effectors.

Fig. S10 Constitutively phosphorylated MoSwa2 cannot rescue the virulent defect of the mutant.

Fig. S11 MoSwa2 regulates the normal localisation of COPII subunit MoSec24‐2.

Fig. S12 MoSec24‐2 is involved in the secretion of Avr‐Pia, Avr‐Piz‐t and Slp1.

Fig. S13 MoSec24‐2 is required for normal virulence and invasive growth.

Fig. S14 Identification and characterisation of extracellular proteins regulated by MoSwa2.

Fig. S15 Function analysis of ROS‐detoxifying enzymes in inhibition of H₂O₂ production and virulence.

Fig. S16 MoSwa2 and MoSec24‐2 are involved in the secretion of MoSef1.

Table S1 Primers used in this study.

Table S2 Vegetable growth and conidiation of wild‐type Guy11, the ΔMoswa2, the ΔMosef1 and the complement mutants.

Table S3 The putative secreted proteins identified in EF from Guy11.

Please note: Wiley Blackwell are not responsible for the content or functionality of any Supporting Information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office.