A novel effector, CsSp1, from Bipolaris sorokiniana, is essential for colonization in wheat and is also involved in triggering host immunity

Abstract The hemibiotrophic pathogen Bipolaris sorokiniana causes root rot, leaf blotching, and black embryos in wheat and barley worldwide, resulting in significant yield and quality reductions. However, the mechanism underlying the host–pathogen interactions between B. sorokiniana and wheat or barley remains unknown. The B. sorokiniana genome encodes a large number of uncharacterized putative effector proteins. In this study, we identified a putative secreted protein, CsSp1, with a classic N‐terminal signal peptide, that is induced during early infection. A split‐marker approach was used to knock out CsSP1 in the Lankao 9‐3 strain. Compared with the wild type, the deletion mutant ∆Cssp1 displayed less radial growth on potato dextrose agar plates and produced fewer spores, and complementary transformation completely restored the phenotype of the deletion mutant to that of the wild type. The pathogenicity of the deletion mutant in wheat was attenuated even though appressoria still penetrated the host. Additionally, the infectious hyphae in the deletion mutant became swollen and exhibited reduced growth in plant cells. The signal peptide of CsSp1 was functionally verified through a yeast YTK12 secretion system. Transient expression of CsSp1 in Nicotiana benthamiana inhibited lesion formation caused by Phytophthora capsici. Moreover, CsSp1 localized in the nucleus and cytoplasm of plant cells. In B. sorokiniana‐infected wheat leaves, the salicylic acid‐regulated genes TaPAL, TaPR1, and TaPR2 were down‐regulated in the ∆Cssp1 strain compared with the wild‐type strain under the same conditions. Therefore, CsSp1 is a virulence effector and is involved in triggering host immunity.


| INTRODUC TI ON
Bipolaris sorokiniana (teleomorph: Cochliobolus sativus) is one of the main pathogens responsible for wheat root rot, crown rot, leaf spot, and black points of wheat, barley, and many other grass species (Acharya et al., 2011;Karov et al., 2009;Kumar et al., 2002Kumar et al., , 2020Yan et al., 2012). Diseases caused by B. sorokiniana have been shown to result in yield losses ranging from 10% to 20% under favourable conditions in Canada, the UK, Brazil, Mexico, Gambia, and southern Asia (Ghazvini & Tekauz, 2012;Kang et al., 2020;Karov et al., 2009;Murray et al., 1998;Sharma & Duveiller, 2010). On the northern plains of China, B. sorokiniana was determined to be the most abundant pathogen present in infected wheat roots and stems (Xu et al., 2018). B. sorokiniana accounts for 0.3%-66.7% of black point disease cases in China, based on screening of wheat cultivars and isolation of the pathogen from grains (Dai et al., 2011;Li et al., 2014;Luan et al., 2011;Zhang et al., 1990). In recent years, with changing climate conditions and returning straw to the field, wheat root rot has become more prevalent in different regions of China. In particular, in the Huanghuai wheat-growing region of China, wheat root rot has become one of the major diseases (Guo, Yao, et al., 2019;Li et al., 2011;Wu et al., 2006;Zhang et al., 2007).

B. sorokiniana is a hemibiotrophic parasitic fungus, and its infection
process is similar to that of Magnaporthe oryzae (Gupta et al., 2018).
The conidia secrete mucus, adhere to the surface of host plants, and then germinate to form buds and germ tubes, which extend and produce multiple branches. Appressoria differentiate at the tops of branches, producing penetration pegs that directly penetrate the cuticle on the surface of host cells, and invasive mycelia extend inside or between host cells (Han et al., 2010;Kumar et al., 2002;Verma et al., 2020). It is rare for hyphae to infect through natural openings such as stomata. Due to heterokaryotic conditions, the morphology of isolates from the field widely vary. In a natural population of B. sorokiniana, the frequency of the greenish-grey colony colour was 31.25%, followed by black (25%) and grey or white (18.75%), whereas brown was the least frequent colour (6.25%) (Verma et al., 2020).
It is well known that salicylic acid (SA) is involved in the plant immune response against biotrophic and hemibiotrophic pathogens, and is associated with the induction of plant systemic acquired resistance (SAR) (Alvarez, 2000;Meenakshi & Singh, 2013;Vlot et al., 2009). SA accumulates during both incompatible and compatible interactions between B. sorokiniana pathogens and host plants to facilitate resistance to spot blotches (Al-Daoude, 2019;Al-Daoude et al., 2018;Sahu et al., 2016). Lesion development is associated with the accumulation of host-encoded pathogenesis-related (PR) proteins and reactive oxygen species (ROS) (Ajith et al., 2003). Among the PR genes, PR1, PR2, and PR5 are commonly used as markers for the activation of SAR . On the other hand, pathogen effectors that suppress SA signalling without affecting SA biosynthesis are also known to exist (Kazan & Lyons, 2014). For example, when expressed in Arabidopsis, two effectors, Hyaloperonospora arabidopsidis effector HaRxL96 and Phytophthora sojae effector PsAvh163, suppress the pathogen-mediated induction of marker genes such as PR1 (Anderson et al., 2012). Many studies have shown that plant pathogens can manipulate SA signalling and that many effectors are involved in this process.
To overcome host detection and defence, most pathogens produce a range of secreted effectors and metabolites (Stergiopoulos & de Wit, 2009;Wit et al., 2010). Two kinds of effectors are found in the cytosol in host cells or the apoplastic space Koeck et al., 2011;Zhang & Xu, 2014). A number of studies have revealed the function of intracellular effectors in filamentous plant pathogens (Bozkurt & Kamoun, 2020;Pramod et al., 2019). In oomycetes, the movement of effectors to host cells occurs via common amino acid sequence motifs, such as RxLR (Arg-x-Leu-Arg), LxLFAK or the Crinkler motif (CRN), and ChxC (Jiang et al., 2008). In fungi, a small group of effectors from barley powdery mildew, wheat stem rust, and wheat leaf rust share a conserved motif, Y/F/WxC, that follows the secreted signal peptide (Godfrey et al., 2010). However, most fungal effectors lack conserved domains (Caillaud et al., 2012;Selin et al., 2016). In M. oryzae, effectors are delivered to the cytosol through a specific structure called the biotrophic interfacial complex (BIC) (Khang et al., 2010). Most of the effectors from pathogens contribute quantitatively to pathogen aggressiveness, but some of them, such as Cmu1 and Scc1 in Ustilago maydis (Djamei et al., 2011;Redkar et al., 2015), BAS107 in M. oryzae , MiSSP7 in Laccaria bicolor, which functions as a negative regulator of jasmonic acid (JA)-induced gene regulation in the nucleus (Plett et al., 2011), PcCRN4 in Phytophthora capsici, which suppresses host defence and induces cell death in the plant nucleus (Mafurah et al., 2015), SsSSVP1 in Sclerotinia sclerotiorum (Lyu et al., 2016), VdSCP7 in Verticillium dahliae , and SCRE1 in Ustilaginoidea virens, which inhibits host immunity and suppresses the immunity-associated hypersensitive response (HR) via the plant nucleus (Zhang, Yang, et al., 2020), have been identified as potential nuclear-localized regulators of the host cell targeting process (Diaz-Granados et al., 2020). The conserved targeting mechanism of a common host protein network for convergent effectors from the eubacteria Pseudomonas syringae, the oomycete Hyaloperonospora arabidopsidis, and the ascomycete Golovinomyces orontii has been explored in the model plant species Arabidopsis thaliana (Weßling et al., 2014). The wheat blue dwarf phytoplasma effector SWP11 induces three PR genes, PR1, PR2, and PR3, to trigger plant immunity . In general, effectors suppress or induce plant cell death mostly through manipulation of the host immune system (Knig et al., 2020;Sharpee & Dean, 2016;Shen et al., 2018;Wang et al., 2011). Some effectors are recognized by the plant immune system through specific resistance proteins and are termed avirulence proteins (Boller & Felix, 2009;Malik et al., 2020). AvrPrm3 in Blumeria graminis is recognized by the Pm3 resistance gene in wheat (Bourras et al., 2015), while Avr2 and Avr3 in Fusarium oxysporum f. sp. lycopersici interact with I-2 and I-3 resistance genes in tomato, respectively (Houterman et al., 2008(Houterman et al., , 2010Rep et al., 2005).
There are few reports on the mechanism underlying the molecular regulation of the pathogenicity of B. sorokiniana in wheat root rot. A comparative analysis of candidate effector-coding genes in the genomes of five Bipolaris species revealed 289 putative smallmolecular-weight secreted proteins in B. sorokiniana, 167 of which were unique. There were significantly more secreted proteins than other pathogens of the same genus, and the functions of these secreted proteins have not been reported (Condon et al., 2013). Pathak et al. (2020 investigated the secretome of 196 proteins predicted to be present in B. sorokiniana in silico. The ToxA gene encodes a hostselective toxin (HST) that functions as an effector, and B. sorokiniana has been shown to carry this gene (Sudhir et al., 2020). However, no experimental evidence has been shown for secreted proteins in B. sorokiniana. In this study, we identified a gene, CsSP1, encoding a small-molecular-weight secreted protein in B. sorokiniana that was highly expressed in the infection stage. This protein is involved in pathogenicity and acts as a novel elicitor triggering the host immune system, and this protein might be a candidate for the control of plant disease.

| CsSp1 is highly expressed during infection
To evaluate gene expression during interactions with wheat, RNA sequencing (RNA-Seq)-based transcriptome analysis was performed using roots and basal stems of Aikang 58 wheat seedlings grown in pots at 5 days and 15 days after soil inoculation with B. sorokiniana (Figure S1a-d). Analysis of transcriptome data revealed very few B. sorokiniana sequence reads. A gene encoding a small protein (9.7 kDa) with a predicted N-terminal signal peptide was highly expressed during the B. sorokiniana infection stage (Figure 1a). We designated this protein as B. sorokiniana-secreted protein 1 (CsSp1).
To verify the expression pattern of CsSp1 during wheat leaf infection, reverse transcription quantitative (RT-qPCR) was conducted.
The results showed that the expression level of CsSP1 increased more than 40-fold at 12 h postinoculation (hpi), decreased at 48 hpi and remained higher than that of hyphae grown in vitro ( Figure 1b). BLASTP analysis of the CsSp1 protein indicated that the amino acid identity ranged from 71.58% to 73.03% with that from Bipolaris maydis, Bipolaris victoriae, and Bipolaris zeicola. Amino acid sequence alignment was performed via DNAMAN (Figure 1c), and a phylogenetic tree was constructed based on the amino acid alignment ( Figure S2c). The results showed that CsSp1 is a protein specific to plant-pathogenic fungi and is found only in the Bipolaris genus.

| Generation of CsSP1 deletion mutants and functional complementation of ∆Cssp1
To characterize the biological function of CsSP1, a split-marker approach was applied to knock out CsSP1 in the wild-type (WT) strain

| CsSp1 is involved in B. sorokiniana conidial regulation
To determine other biological functions of CsSp1, we counted the number of spores present on PDA plates. Compared with the WT, ∆Cssp1 exhibited less sporulation (Figure 3a,b) and, in terms of morphology, the spores were smaller than those of the WT (Figure 3d).
To explain the decrease in spore production, we measured the expression levels of the orthologous genes CsBrlA, CsMedA, and CsStuA (Wang et al., 2015), which are essential for the positive regulation of sporulation in both mutant and WT B. sorokiniana. The results showed that the expression levels of these candidate genes significantly decreased after 24 h of cultivation ( Figure 3c). Taken together, these results confirmed the previous results in which CsSp1 is involved in the regulation of spore formation in B. sorokiniana.

| CsSp1 is a virulence factor needed for full virulence of B. sorokiniana
To evaluate the role of ∆Cssp1 in pathogenesis, we subjected wheat seedlings to a soil inoculation assay involving the application of 5-mm diameter fungal agar plugs to the stem bases. The pathogenicity test indicated that the virulence to wheat rot of ∆Cssp1 was nearly completely lost ( Figure S3a). The leaves were inoculated with spore suspensions (3 × 10 4 spores/ml), and the results showed that the ∆Cssp1 mutants caused only tiny black spots on the leaves, while the WT pro-  (Figure 4d). The spores germinated normally; however, the morphology of the hyphal tips was altered, and the tips changed direction. ∆Cssp1 displayed abnormal curving or swelling that seemed to initiate appressorium differentiation that had failed. The proportion of normal appressoria in ∆Cssp1 was significantly lower than that in the WT ( Figure S3d). Compared with the WT hyphae, the invasive ∆Cssp1 hyphae were swollen and stunted, and extended more slowly (Figure 4h). Therefore, the CsSP1 gene encodes a virulence factor involved in vegetative development and infection structure.

| CsSp1 is a secreted effector that localizes to the nucleus and cytoplasm of host cells
Further bioinformatics analysis indicated that the protein encoded by CsSP1 was smaller than 10 kDa and lacked similarity to proteins F I G U R E 1 Differentially expressed genes (DEGs) in the transcriptome and CsSP1 feature analysis. (a) Cluster analysis of DEGs from the wheat Aikang 58 cultivar infected with Bipolaris sorokiniana. (b) Reverse transcription quantitative PCR (RT-qPCR) confirmation of the expression pattern of CsSP1 in planta. Aikang 58 leaves were inoculated with 10 5 spores/ml of B. sorokiniana Lankao 9-3. The inoculated leaves were placed in a moist chamber in the dark for 24 h and then kept in a greenhouse at 25°C (47% humidity) with a 16 h light/8 h dark photoperiod. The leaves were sampled at 12, 24, 36, and 48 h for total RNA extraction, while mycelia cultured for 2 days in YEPD were used for fungal RNA extraction. The experiments were repeated three times. The expression levels were calculated using the 2 −∆∆Ct . Significant differences calculated by Tukey's LSD, p < 0.05.  Table 1).
To verify the function of the predicted signal peptide of CsSP1, the DNA fragment encoding the signal peptide was introduced into a pSUC2 vector (yielding a pSUC2-CsSP1-SP construct), which was subsequently transformed into yeast strain YTK12 to examine F I G U R E 2 CsSP1 knockout and characterization. (a) Diagram of the gene deletion strategy for CsSP1. The primers used for gene replacement and mutant screening are indicated by the arrows and are listed in Table 2. (b) Agarose gel electrophoresis of PCR products from genomic DNA templates. Lane H, primers HYG-F/HYG-R for hygromycin resistance gene; lane F and lane R, CsSP1-PF/H855R and H856F/CsSP1-PR, respectively, for positive screening; lane G, CsSP1-NF/CsSP1-NR for negative screening. ∆Cssp1-3 and ∆Cssp1-4 are two candidates: wild-type (WT) strain Lankao 9-7; M, molecular markers; H, hygromycin resistance gene; F, upstream; R, downstream; G, CsSP1 gene. (c, d) The colony morphology and growth rate calculated for 90 mm potato dextrose agar (PDA) plates after 7 days. The bars indicate the standard errors. The experiments were repeated three times. Significant differences calculated by Tukey's LSD, p < 0.05. (e) Morphology of hyphal tips on PDA plate. Bars, 50 μm secreted invertase activity. Avr1b-SP was used as a positive control.
Growth tests on synthetic tryptophan (Trp) dropout agar medium plates showed that CsSP1-SP and Avr1b-SP restored the secretion of invertase and resulted in yeast growth on sucrose medium ( Figure 5b). The activity of secreted invertase was also measured by the reduction of 2,3,5-triphenyltetrazolium chloride (TTC), and the secreted invertase of the transformant containing Avr1b-SP and CsSP1-SP was measured by TTC assays (Figure 5b). CsSP1-SP restored invertase such that TTC became red insoluble triphenylformazan (TTF) (Figure 5b). Thus, the CsSp1 signal peptide is functional.
To determine the subcellular localization of CsSp1, a C-terminal green fluorescent protein (GFP) fused to CsSP1 was used. CsSP1 was cloned and ligated into a PVX-GFP-3HA plant expression vector, which was then transiently expressed through agro-infiltration

| In planta expression of CsSp1 suppresses pathogen extension
CsSP1 was amplified via PCR and cloned into a vector PB-3HA plant expression. Agro-infiltration of PB-CsSP1 did not induce cell death (Figure 5f). Because B. sorokiniana itself does not infect F I G U R E 3 ∆Cssp1 spore morphology, conidiation, and potential regulation in the wild type (WT) and complemented (cCssp1) strains.

| D ISCUSS I ON
Secreted effector proteins play indispensable roles in interactions between plants and phytopathogenic fungi (Giraldo & Valent, 2013). Orthologous genes of CsSp1 may play important roles in Bipolaris species. The pathogenicity-related effector protein in S. sclerotiorum, SsCP1, was also expressed at the highest level at 12 hpi and remained at a high level for at least 48 hpi. S. sclerotiorum SsCP1 induces cell death via the host immune response, while CsSp1 does not cause cell death .
In Metarhizium acridum, deletion of MaPMT1 does not affect appressorium formation but significantly decreases appressorium turgor pressure to weaken virulence (Wen et al., 2020). In M. oryzae, mutant strains lacking the spermine synthase-encoding gene SPS1 progress through all stages of appressorial development, including penetration peg formation, but cuticle penetration is unsuccessful due to reduced appressorial adhesion, which leads to solute leakage (Rocha et al., 2020). ∆Cssp1 formed an abnormal appressorium, and infectious hyphal swelling indicated restricted extension in plant cells. Branching vegetative hyphae also exhibited swelling on PDA plates, and bent tip hyphae occurred on onion epidermal cells. On PDA, B. sorokiniana usually easily produces abundant spores (Guo, 2016). We observed a decrease in spore production by ∆Cssp1.
Despite the decreased size of the ∆Cssp1 spores, they could still germinate. The profile of sporulation is highly regulated, and most of the key regulators are conserved throughout filamentous fungi (Chung et al., 2011;Park & Yu, 2012;Zhao et al., 2015). We identified three key regulators, the orthologous genes CsBrlA, CsMedA, and CsStuA, in B. sorokiniana, and their expression levels significantly decreased at 24 h after cultivation. Therefore, CsSp1 is essential for vegetative development and asexual reproduction.  (Dong et al., 2015), Colletotrichum orbiculare (Yoshino et al., 2012), Fusarium oxysporum (Gawehns et al., 2014), and Valsa mali  have also been reported. Nonetheless, the mechanism by which B. sorokiniana suppresses the plant immune response needs to be determined. were not the same as the predicted ones shown in Figure 5a. We used YinOYang 1.2 Server (http://www.cbs.dtu.dk/servi ces/yinoy ang/) to predict O-glycosylation sites, six of which were identified in CsSp1. N16 was predicted to be a glycosylation site by YinOYang 1.2 Server (Table 1). In our study, N16 was deleted as part of the signal peptide. Removing two additional amino acid residues (in which one is a potential glycosylation site) might cause the mature protein to be unstable. N-glycosylation shields the P. sojae apoplastic effector PsXEG1, which is sensitive to a specific host aspartic protease (Xia et al., 2020), so we hypothesize that glycosylation of CsSp1 may also affect protein degradation by some proteases. This did not affect the Plant hormones, including the classic ones SA, JA, and ethylene (ET), regulate the defence response to various pathogens (Kazan & Lyons, 2014). SA is essential for modulating the regulatory response to biotrophic and hemibiotrophic pathogens (Han & Kahmann, 2019).
It is widely known that B. sorokiniana mainly triggers the SA signalling pathway, but the underlying mechanism remains unknown (Aldaoude, 2019). In our study, compared with uninoculated leaves, B. sorokinianainoculated leaves (at the early stage) presented a substantial amount of PAL activity (Mali et al., 2017;Singh et al., 2019). This is consistent with our findings that the expression of TaPAL was up-regulated in the infection stage and that the expression of TaPAL in ∆Cssp1 cells was lower than that in WT cells. In S. sclerotiorum, the virulence factor SsCP1 indeed interacts directly with the PR protein PR1 in planta . Moreover, SA-dependent genes including PR1, PR2, and Collectively, this study provides information on some functional features of CsSp1, paving the way for an improved understanding of the molecular mechanism through which pathogen effectors manipulate the plant immune system.

| Plant materials and fungal strains
The susceptible wheat variety Aikang 58 was used for bioas-

| Gene knockout
The split-marker gene knockout strategy was used in this study according to the description of Wang et al. (2017), with slight modifications. Briefly, for homologous recombination fragments, upstream and downstream fragments of the target gene (approximately 1 kb) were amplified from the genomic DNA of B. sorokiniana Lankao F I G U R E 5 Functional identification of CsSp1 secretion, translocation, and elicitor characteristics. (a) The signal peptide of CsSp1 between 17 and 18 amino acid residues was predicted by SignalP. (b) Validation of signal peptides in a yeast system. If SUC2 invertase is secreted extracellularly in yeast, TTC can be reduced to red TTF. CsSp1, CsSp1 signal peptide with the expression vector pSUC2-CsSP1-SP in yeast; YTK12, empty yeast; Avr1b, avirulence gene b signal peptide, which served as a positive control. (c) Subcellular localization of CsSP1-GFP in Nicotiana benthamiana. (d) Western blot analysis confirming protein expression with the PVX-HA vector with an HA tag using protein from N. benthamiana leaves 48 h after Agrobacterium tumefaciens injection and leaves treated at 95°C for 5 min before protein extraction. CsSP1-GFP and PVX-CsSP1-GFP include signal peptides. Mouse anti-HA antibodies (M20003) and goat anti-mouse secondary antibodies (IgG horseradish peroxidase conjugate; M21001) with working concentration of 1:5000 from Abmart were used (http://www. ab-mart.com/). CsSP1-NS, PVX-CsSP1-NS-GFP lacking a signal peptide. GFP, PVX-GFP. (e) Subcellular localization of CsSP1-GFP in onion epidermis cells. The constructs used were the same as those in (d). The onion epidermis was treated with the corresponding Agrobacterium.  Split hygromycin B resistance gene fragments were generated from pKOV21 plasmid DNA. The three fragments were fused together via PCR homologous recombination. Next, using the fusion fragment as the template, the upstream region of the target gene and the front two-thirds of the hygromycin resistance gene sequence were amplified, after which the back two-thirds of the hygromycin resistance gene sequence and the downstream region of the target gene were amplified. Two fragments were prepared for protoplast transformation. The amplified fragments were purified using B518141-0100 (SanPrep Column PCR Product Purification Kit) from Sangon Biotech.
The protoplast preparation methods were based on those for Fusarium graminearum .

| Fungal growth and sporulation
The spores of the original WT strain Lankao 9-3 in glycerol were kept in a freezer at −80°C. Cultures for fungal growth, sporulation, and pathogenicity tests started from this stock. A 5-mm fungal block was shredded with a blender and then placed on 15 ml of PDA (200 g peeled potato, 20 g dextrose, 20 g agar, 1 L water), which was placed in an incubator at 25°C under darkness for 7 days. The diameter of colony growth was measured after 7 days.
The hyphae were then removed and cultured under light for 5 days, A spore solution (20 μl, 5 × 10 4 /ml) was dropped onto the onion epidermis. The cells were placed in a 25°C illuminated humid chamber, and the structure of the appressoria was observed via microscopy after inoculation.
For data analysis, the standard deviation was used for all errors throughout the experiments. Significant differences were calculated by Tukey's LSD, p < 0.05 and t test.

| RNA-Seq and data analysis
The details of Aikang 58 wheat seedlings infected with or without B. sorokiniana WT strain Lankao 9-3 in pots were described by Kang et al. (2020). Seedling samples were carefully collected at 5 and 15 days after inoculation and rinsed with tap water to assess disease development. The same method was used to harvest samples from the field. The collected clean stem bases and roots were subsequently frozen in liquid nitrogen and kept at −80°C until use. We

| Construction of expression vectors
For complementation of the knockout mutant, the target gene CsSP1 (e, f) RT-qPCR for TaPR1 and TaPR2, respectively, of wheat leaves infected with spores. Significant differences of samples with the same spore infection calculated by Tukey's LSD, p < 0.05. The lower case letters refer to a significant difference between the wild-type (WT) samples and upper case letter refers to a significant difference between the ∆Cssp1 samples. The significance between samples of the same infection period was tested by t test. **p < 0.01, *p < 0.05 (t test). The experiments were repeated three times. The expression levels were calculated using the 2 −∆∆Ct method PstⅠ sites at the 5ʹ and 3ʹ ends. The fragment amplicons were subsequently cloned and ligated into a pYIP-102 expression vector at the

| Polyacrylamide gel electrophoresis and western blotting
For protein extraction, mycelial or plant samples were ground to a fine powder in liquid nitrogen. The fine powder was then treated with protein extraction buffer (20 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM EDTA, 20% glycerol, 1 mM phenylmethylsulfonyl fluoride [PMSF], and a suitable concentration of protease inhibitor). After swirling and mixing, the proteins were released after 10 min in an ice bath. After centrifugation at 13,000 × g for 15 min, the supernatant, which contained the desired protein, was transferred to a new centrifuge tube. The mixture was brought to 5 ml with double deionized water and mixed well, with β-mercaptoethanol added just before use. After immersion in boiling water for 10 min, the sample was used for protein extraction.
For western blotting, the methods were performed according to those of Li et al. (2018), and anti-HA antibodies (M20003) were used (http://www.ab-mart.com/).

| Yeast secretion assays
A pSUC2 vector containing a tryptophan synthesis-related gene but lacking a signal peptide, as well as the invertase gene (SUC2) without the start codon (ATG), was used. The product of the invertase gene (SUC2) converts polysaccharides to monosaccharides. Therefore, only when the secreted gene is inserted can the missing SUC2 gene be activated and secreted into the culture medium for the conversion of sucrose to glucose needed for yeast growth. The methods used refer mainly to the protocols described previously (Gu et al., 2011;Li et al., 2016Li et al., , 2019. We integrated these approaches and made several modifications. First, 10% dimethyl sulphoxide (DMSO) was added to the yeast competent cells. In addition, sucrose selective medium (SD−Trp−sucrose medium, 0.8 g yeast synthetic drop-out medium without Trp, 2% sucrose, 2% agar) was used for the final screening (Guo, Zhong, et al., 2019).

| Gene expression analysis
Total RNA for measuring CsSP1 expression was extracted from tively, was measured (Wang et al., 2015).
All the internal reference primers used to express the wheat genes were Actin-PR-F/Actin-PR-R (Naz et al., 2018). The primers used to detect TaPAL were TaTAL-F/TaPAL-R, while those used to detect TaPR1 and TaPR2 were TaPR1-F/TaPR1-R and TaPR2-F/ TaPR2-R (Niu et al., 2007). All the primers used are listed in Table 2.
The relative transcript levels of test genes were determined according to the function ∆C t = C t (test gene) − C t (reference gene). Briefly, the threshold cycles (C t ) of the PCR results for each gene were first obtained and then averaged for use of quantification of the transcripts in the next step. The ∆C t value was determined by subtracting the average C t value of the endogenous reference genes, Actin or EF1α in this study, from the average C t value of the candidate gene.
The ∆∆C t value was calculated by subtracting the ∆C t value of the B. sorokiniana mycelia control from the ∆C t of the inoculated sample.
The 2 −∆∆Ct value was used to evaluate the fold change of gene expression. The relative expression under different conditions was calculated according to the 2 −ΔΔCt method (Livak & Schmittgen, 2001).
Each experiment had three replicates to give the main value and the standard deviations were generated.

ACK N OWLED G EM ENTS
We would like to thank Dr Yuanchao Wang from Nanjing Agricultural

DATA AVA I L A B I L I T Y S TAT E M E N T
The RNA-Seq data are available from GenBank at https://www.ncbi.

S U PP O RTI N G I N FO R M ATI O N
Additional supporting information may be found in the online version of the article at the publisher's website.