Amino acid sequences of Homo sapiens-, Fusarium graminearum-, and Neurospora discreta-annotated eIF3k were used to run BLASTp and reverse BLASTp research on the fungi and Oomycetes genomic resources platform (Basenko et al., 2018). Results obtained from these multiple blast searches identified a single copy of an uncharacterized (hypothetical) protein containing CSN8/PSMD8/EIF3K domain, here referred to as (MoOeIF3k) in the genome sequence assembly data available for the rice blast fungus. Additional phylogenetic analyses were performed to determine the evolutionary pattern of eIF3ks at both class and kingdom levels using MEGAX, Japan version 10.1.7 (Molecular Evolutionary Genetics Analysis software) (Kumar et al., 2016; Barazesh et al., 2020) and deployed online on the Interactive Tree Of Life (iTOL) platform (Letunic and Bork, 2007) for final polishing. Results from these examinations revealed that within the class Sordariomycetes, MoOeIF3k shared a common evolutionary clade with eIF3ks identified in Sporothrix schenckii, Sporothrix brasiliensis, Sordaria macrospora k-hell, Neurospora tetrasperma, and N. discreta; MoOeIF3k is, however, evolutionarily distant from eIF3ks identified in F. graminearum, Fusarium oxysporum, Fusarium proliferatum, Fusarium verticillioides, Fusarium fujikuroi, Trichoderma virens, Trichoderma reesei, Lomentospora prolificans, and Scedosporium apiospermum. Meanwhile, eIF3ks from fungi populations in the class Sordariomycetes clustered into a single heterogeneous clade but are phylogenetically distant from eIF3ks from members in the class Eurotiomycetes (Aspergillus fumigatus) and Leotiomycetes (Botrytis cinerea and Sclerotinia sclerotiorum). We also observed that eIF3ks from Sordariomycetes are phylogenetically distant from plants (Arabidopsis thaliana and Oryza sativa japonica), mammals (H. sapiens), Oomycetes, and other fungi classes, such as Saccharomycetes, Ustilaginomycetes, Glomeromycetes, Blastocladiomycetes, and others, lack eIF3k orthologs in their genomes (Figure 1A).

Additional results acquired from Pfam 32.0 domain profiling analysis (El-Gebali et al., 2019) showed that notwithstanding the within and between class and kingdom level phylogenetic differences recorded among eIF3ks from the selected organisms, all the eIF3k sequences from fungi, plants, and mammals possess the typical CSN8/PSMD8/EIF3K (Figure 1A).

To determine the functional role of MoeIF3k in the development of M. oryzae Oryza (MoO), we accordingly deployed homologous recombination strategies and generated targeted gene deletion strains (ΔMoOeif3k), and complemented the ΔMoOeif3k strains by transforming the full-length open reading frame (ORF) MoOeIF3K and MoOeIF3K-GFP fusion constructs into protoplasts prepared from the ΔMoOeif3k strains independently and obtained two lines of complementation strains, ΔMoOeif3k_Com-1 and ΔMoOeif3k_Com-11. Results from Southern blotting assays revealed the successful replacement of the MoOeIF3K gene by hygromycin B resistant gene at a single locus. Meanwhile, one of the multiple bands observed in the Southern blotting images obtained for the ΔMoOeif3k_Com-1 and ΔMoOeif3k_Com-11 strains corresponds to the 3.6 kb band in ΔMoOeif3k (Figures 1B,C). The results indicated that the complementation strains obtained in this study resulted from random and multiple insertions of the MoOeIF3K and MoOeIF3K-GFP constructs at other loci within the genome without displacing the hygromycin B-resistant gene from the ΔMoOeif3k_Com-1 and ΔMoOeif3k_Com-11 strains. These results suggest that MoeIF3k is a member of the MoO non-essential eIF3 complex and plays a functionally dispensable role in the survival of the rice blast fungus.

To gain insights into the contributions of MoOeIF3k to the vegetative development of MoO, the ΔMoOeif3k strains generated in this study were cultured on different types of nutrient-sufficient and nutrient-deficient culture media. Growth assessment records showed a significant decrease in the vegetative development of ΔMoOeif3k strains cultured on different kinds of nutrient-sufficient media, including complete media, straw decoction, and corn agar medium (SDC), prune agar (PA), and rice bran agar (RBA) (Figures 2A,B). Conversely, there was a non-significant but appreciable increase in the vegetative growth of ΔMoOeif3k strains grown on nutrient-deficient media, namely, minimum media (MM) and water agar (WA) media. Meanwhile, MM and WA supplemented with 2% w/v glucose as a carbon source triggered a substantial reduction in the growth of ΔMoOeif3k strains (Figures 2C,D). Since translation is an energy- and nutrient-intensive cellular process (Doudna and Sarnow, 2007), we averred that the absence of MoOeIF3k would not only trigger a reduction in the translational load (protein turn-over) but could also alter the translational dynamics to favor the translation of proteins essential for the mitigation of stress homeostasis in MoO. Accordingly, we posited that MoOeIF3k acts as a positive and negative regulator of fungal morphogenesis under nutrient-sufficient and deficient conditions, respectively, through direct or indirect translational regulation of proteins associated with morphological development in filamentous fungus under nutrient-specific conditions.

In C. elegans, eIF3k functions as a negative regulator of longevity and stress tolerance (Cattie et al., 2016). Therefore, to examine the impact of MoOeFI3K gene deletion on the response of the rice blast fungus to oxidative and reductive stress, the ΔMoOeif3k strains, along with the complementation and wild-type strains, were cultured on CM supplemented independently with Calcofluor White (CFW), NaCl, SDS, and Congo red (CR) as oxidative (Just and Arendshorst, 2003; Ding et al., 2014; Messina et al., 2014; Khare et al., 2015) stress-inducing agents and DTT, tunicamycin (Tu), and thapsigargin (Tg) as reductive stress-inducing agents (Grant, 2002; De Minicis et al., 2012; Korge et al., 2015; Arrieta et al., 2020; Ma et al., 2020). Results obtained from stress-responsive bioassays showed that the growth ΔMoOeif3k strains on complete media (CM) supplemented with oxidative stress-inducing agents, such as sodium dodecyl sulfate (SDS) and NaCl, were significantly inhibited. We observed that the ΔMoOeif3k strains were more resistant to CR than the wild-type and complementation strains (Figures 3A,B). The growth of the individual strains on MM supplemented with oxidative and reductive (ER) stress-inducing osmolytes yielded a similar sensitivity response (Supplementary Figures 1A,B). The ΔMoOeif3k strains are significantly sensitive to Tu but immune to DTT and Tg, as well as the mock control (DMSO) (Figures 3C,D and Supplementary Figures 1C,D). Meanwhile, results obtained from temperature sensitivity assays showed that the ΔMoOeif3k strains displayed almost stable vegetative growth characteristics as the wild-type and complementation strains (Supplementary Figures 2A,B).

Further quantitative PCR (qPCR) evaluation of the expression pattern of the MoOeFI3K gene during oxidative and reductive stress response of the rice blast showed that the exposure of MoO to different stress-inducing conditions did substantially alter the transcription pattern of MoOeFI3K (Figure 3E). However, Western blotting bio-assays revealed a reduction in the expression level of MoOeIF3k exclusively during DTT, CR, and UV-induced stress conditions (Figure 3F). On the contrary, we demonstrated that the exposure of MoO to Tu- and Tg-induced ER stress (reductive stress) enhanced the accumulation of MoOeIF3k (Figure 3G). Meanwhile, we also observed that the deletion of MoOeIF3K triggered a slight but non-significant increase in the expression pattern of other MoOeIF3 subunits, especially MoOeIF3D and MoOeIF3H (Supplementary Figure 3). We subsequently inferred that MoOeIF3k contributes significantly to the regulation of osmolyte-specific stress tolerance of MoO.

To unravel the impact of MoOeIF3K gene deletion on translation and post-translational processes, we performed a comparative quantification of total ribosomal RNAs (rRNAs) and total protein turnover between the wild-type and defective strains. About 80% of cellular RNAs are rRNAs. Therefore, an rRNA quantification analysis was performed by measuring the concentration of total RNAs extracted from the individual strains and used to compute the approximate total rRNA content using the algorithm (Estimated total rRNA = Concentration of total RNA × 0.8). A comparative analysis of estimated total rRNA level between the wild-type and ΔMoOeif3k strains showed that targeted disruption of the MoOeIF3K subunit triggered a substantial increase in rRNA generation in the defective strains (Table 1).

Also, to ascertain the significance of the substantial increase in rRNA level on total protein output in the ΔMoOeif3k strains, we performed a bicinchoninic acid (BCA) assay-mediated comparative quantification of total protein turnover between the wild-type and ΔMoOeif3k strains. The examinations revealed an increased total protein level in the ΔMoOeif3k strains compared with the wild-type (Table 1). Given the concomitant increase in total rRNA and protein level in the MoOeIF3K-defective strain, we concluded that MoOeIF3K likely influences multiple developmental processes in the rice blast fungus through the regulation of protein synthesis.

Sporulation plays a significant role in both the perpetuation and dissemination of the rice blast disease. Unique proteins or pathways drive asexual and sexual reproduction in fungi (Cvitanich and Judelson, 2003; Schoustra et al., 2010; Park and Yu, 2012; Son et al., 2014). To unravel the potential contributions of MoOeIF3k to asexual reproduction in the rice blast fungus, we monitored conidiophore formation. We also performed a comparative assessment of the number of conidia produced by the ΔMoOeif3k, complementation, and wild-type strains cultured on RBA to induce the production of asexual spores. We observed that the deletion of MoOeIF3K attenuated conidiophore development and significantly suppressed conidia production in the ΔMoOeif3k strains (Figures 4A–C).

Additional results obtained from mating assays performed in this study to examine the influence of MoOeIF3K on sexual reproduction in MoO revealed that MoOeIF3K modulates the progression of sexual reproduction in the rice blast fungus independent of MoOeIF3k function. Hence, sexual sporulation characteristics displayed by the ΔMoOeif3k strains were indistinguishable from those of the wild-type and complementation strains (Figure 4D). These results suggest that MoOeIF3k likely influences asexual reproduction in MoO by modulating cellular processes or pathways independent of sexual reproduction in filamentous fungi.

To examine the pathogenicity or virulence efficiency of hyphae and asexual spores produced by ΔMoOeif3k, we inoculated intact and injured leaves of the “golden promise” barely cultivar with mycelia harvested from the ΔMoOeif3k strains, along with the complementation and wild-type strains. At the same time, barley leaves were drop-inoculated with spores in suspension, and leaves of the susceptible rice seedlings (CO39) were spray-inoculated with spores from the individual strains. Records obtained from these infection assays showed that the deletion of MoOeIF3K severely compromised both mycelia, and that conidia mediated the development of blast symptoms on both intact and injured leaf tissues (Figures 5A–C). Additional results obtained from a semi-quantitative assessment of virulence characteristics of the individual strains revealed a remarkable reduction in the severity of blast infections caused by the ΔMoOeif3k strains on the blast-susceptible CO39 rice seedlings (Figure 5D).

Appressorium and hyphae tip appressorium-like structures are the fundamental infectious structures that support the development of conidia- and hyphae-mediated blast infections, respectively. To identify some of the possible factors accounting for attenuation in the virulence of the ΔMoOeif3k strains, conidia produced by the ΔMoOeif3k strains, the wild-type strain, and the ΔMoOeif3k_Com strain were inoculated on appressorium-inducing hydrophobic coverslips to assay the germination and appressorium formation characteristics of the ΔMoOeif3k strain relative to the wild-type and complementation strains in vitro. The formation of hyphae tip appressorium-like structures was assayed by inoculating mycelium from the ΔMoOeif3k strain, the wild-type strain, and the ΔMoOeif3k_Com strain on the barley leaves and on hydrophobic coverslips (Lin et al., 2019). Corresponding results obtained from microscopy examinations showed that the morphology of the appressorium produced by the ΔMoOeif3k strains was comparable to those produced by the wild-type strain (Figures 5E–H).

However, results obtained from histopathological examinations showed that targeted gene disruption of MoOeIF3K severely compromised both penetration and colonization efficiencies of ΔMoOeif3k (Figures 5I,J). We inferred that virulence defects displayed by the ΔMoOeif3k strains are not due to distortions in the morphological integrity of appressoria and appressorium-like structure formed by the ΔMoOeif3k strains.

In an attempt to evaluate the functionality of appressorium and appressorium-like structures formed by the ΔMoOeif3k strains, we instituted glycerol-mediated incipient cytorrhysis assays to monitor the accumulation of turgor pressure in the appressorium and appressorium-like structures. Findings from these investigations revealed that targeted gene disruption of MoOeIF3K compromised appressorium and appressorium-like structure turgor (Figures 6A,B).

We monitored the integrity of appressorium produced by the individual strains by independently staining with iodine and BODIPY^TM 493/503 at different stages to observe the mobilization and accumulation of glycogen and lipid bodies, respectively.

Furthermore, we demonstrated that the deletion of MoOeIF3K substantially delayed the timely mobilization of glycogen to the appressorium of ΔMoOeif3k strains 8, 18, and 24 h post inoculation (hpi) relative to the wild-type and complementation strains (Figures 6C–E). Interestingly, we noticed that the deletion of MoOeIF3K had no adverse effect on the mobilization of lipid bodies during appressorium morphogenesis in MoO; hence, lipid body mobilization features displayed by the ΔMoOeif3k strains were comparable with those displayed by the wild-type and complementation strains (Supplementary Figure 4).

Accordingly, we concluded that MoOeIF3k likely contributes to appressorium integrity through selective translational regulation of proteins associated with turgor development and glycogen mobilization.

From the fundamental knowledge that translation events occur on either free ribosomes within the cytosol or ER-associated ribosomes, we envisage that eIFs could localize to either the cytoplasmic region or the ER. However, results obtained from previous subcellular localization assays showed that eIFs, such as eIF3k, localize to multiple cellular compartments, such as nuclear (Ahmed, 2001; Marchione et al., 2013). Therefore, to ascertain the subcellular localization of MoOeIF3k at different developmental stages of MoO, we accordingly performed microscopy assays using hyphae, conidia, and appressorium obtained from the ΔMoOeif3k_Com strain harboring a MoOeIF3k-GFP fusion construct. The MoOeIF3k-GFP displayed a filamentous cytoplasmic localization pattern with intense GFP fluorescence signals around the perinuclear region (Figures 7A–C). Meanwhile, MoOeIF3k-GFP and Histone-mCherry (His-mCherry) fusion co-localization assays confirmed that MoOeIF3k-GFP localizes to both the cytoplasm, and the perinuclear region at all developmental stages of the rice blast fungus but not the nuclear (Figures 7D–F). At the same time, we demonstrated that MoOeIF3k-GFP co-localized with the ER marker (Kar2-mChery) to the ER at all developmental stages (Figures 7G–J). These results indicated that MoeIF3k mediates the translation of mRNAs in the cytoplasm and the ER in MoO.

To identify a possible cross-talk between MoOeIF3k and transcriptional regulation machinery during pathogen–host interaction (PHI), we extracted protein Guy 11 strains harboring MoOeIF3k-GFP fusion constructs and control Guy 11 strains containing the vehicle vector with GFP (RG7) for co-immunoprecipitation (Co-IP) assays using GFP-Trap beads. Comparative co-IP analyses revealed a total of 531 MoOeIF3K-GFP co-immunoprecipitation protein complexes. Furthermore, Gene Ontology (GO) and pathway enrichment analyses showed that putative MoOeIF3k-interacting proteins recovered from the MoOeIF3k-GFP immune-complex predominantly consist of proteins associated with, ribonucleoprotein complex (171 proteins), rRNA processing (93 proteins), and regulation of gene expression (215 proteins) (Figure 8A). Putative MoOeIF3k interacting proteins associated with the regulation of gene expression include 34 transcription factors (TFs) (Supplementary Table 2) and other components of the eIF3-complex.

Also, we examined the likely co-regulation of MoeIF3k on the 34 TFs recovered from the MoOeIF3k-GFP immune-complex during PHI by monitoring the expression pattern of MoOeIF3K along with the 34 TFs in multiple infection-stage transcriptomic data generated by Jeon et al. (2020) and deposited at public repository database NCBI Sequence Read Archive and assigned with accession numbers SRR8259726 to SRR8259732. Results obtained from these examinations revealed a partial co-regulation and induction in the expression of MoOeIF3K and putative TFs encoded by DNA-directed MGG_03215 (MoORPB3), MGG_02493 (MoOTF2), and MGG_07728 (MoOTF1) (Figure 8B). Additional RT-qPCR assays confirmed a partial correlation between the expression of MoOeIF3K, MoORPB3, MoOTF2, and MoOTF1 during PHI (Figure 8C). These observations suggest a possible co-regulation cross-talk between transcriptional and translational regulatory machinery during the progression of PHI.

Also, co-IP-mediated interactome analyses revealed MoOCsn4 as one of the putative interactors of MoOeIF3k (Supplementary Table 3). Since earlier domain search analyses identified a conserved CSN8_PSD8_eIF3K motif as an inseparable domain across corresponding orthologs from animals, fungi, and plants species sampled for phylogenetic analyses conducted in this study, we deployed yeast-two-hybrid (Y2H) assays to examine the possible existence of direct/physical interaction of MoOeIF3K and subunits of the COP9/CSN-signalosome complex in the rice blast fungus and also to partly dispel any likely assumptions that might seek to project MoOeIF3K as a subunit of the CSN complex. Interestingly, while no physical interaction was observed between MoOeIF3k-BD vs. MoOCsn1-AD, MoOeIF3k-BD vs. MoOCsn2-AD, MoOeIF3k-BD vs. MoOCsn3-AD, MoOeIF3k-BD vs. MoOCsn4-AD, MoOeIF3k-BD vs. MoOCsn6-AD, MoOeIF3k-BD vs. MoOCsn7-AD, and MoOeIF3k-BD vs. MoOCsn12-AD, our investigations revealed the existence of a possible physical interaction between MoOeIF3k-BD vs. MoOCsn5-AD (Figures 9A,B). However, previous studies have shown that Csn5-AD interacts directly with the GAL4 DNA-binding domain (DBD) to yield false-positive results in GAL4-mediated Y2H assays (Nordgård et al., 2001). Hence, to confirm whether the interaction observed between MoOeIF3k-BD vs. MoOCsn5-AD emanated from false binding between the GAL4-DBD domain and MoOCsn5-AD, about 10 μl of AH109 yeast strains harboring the empty GAL4-DBD (pGBKT7) and MoOCsn5-AD constructs were grown on SD-Leu-Trp-His-Ade media with surface fortified with X-α-gal (Supplementary Figure 5). Results from these examinations confirm the interaction between the DB domain-containing vehicle pGBKT7 and MoOCsn5-AD. Meanwhile, no interaction was observed between MoOCsn5-AD and pGADT7 when strains harboring MoOCsn5-AD pGADT7 constructs were cultured on SD-Leu-Trp-His-Ade media with the surface fortified with X-α-gal. Additional results from bimolecular fluorescence complementation (BiFC) analyses performed in this study further confirmed the absence of direct interaction between MoOeIF3k and MoOCsn5 (either MoOCsn5-CYFP + MoOeIF3k-NYFP, MoOCsn5-CYFP + NYFP, or CYFP + MoOeIF3k-NYFP) in vivo (Figure 9C). The absence of physical interaction between MoOeIF3K and the eight subunits of the CSN complex in MoO coupled with the high sequence similarity of MoOeIF3K with eIF3K orthologs from other organisms further confirmed MoOeIF3k as a putative member of the eIF3 complex.

The parental wild-type MoO (Guy11) strain used as a background in the generation-targeted gene-replacement mutant strains for the MoOeIF3K gene characterized in this study was a gift from Dr. Didier Tharreau (CIRAD, Montpellier, France). Bacteria competent cells used to propagate the constructed plasmids were prepared from Escherichia coli strain DH5α.

For vegetative growth (either assessment of vegetative growth), the wild-type, mutant, and complementation strains were cultured on a complete medium (CM, 6 g yeast extract, 6 g casein hydrolysate, 10 g sucrose, 20 g agar), PA medium (40 ml prune juice, 2.5 g lactose, 2.5 g sucrose, 1 g yeast extract, 20 g agar, and pH 6.5), SDC (100 g of straw, 40 g of corn powder, and 15 g agar), MM (6 g NaNO₃, 0.52 g KCl, 0.152 g MgSO₄⋅7H₂O, 1.52 g KH₂PO₄, 0.01 g VB1, 1 ml microelement, and 20 g agar), WA (20 g agar) at 25°C. For conidiation, the strains were cultured on RBA medium (40 g rice bran, 20 g agar, and pH 6) for 10 days in the dark. The plates were later transferred into an incubator with continuous light for 3 days after removing the vegetative hyphae. For sensitivity assays, the individual strains were cultured on CM plates supplemented with different stress-inducing agents (oxidative stress-inducing agents): 200 μg/ml CFW (F3543; SIGMA, Germany), 0.7 M NaCl, 0.01% sodium dodecyl sulfate, 200 μg/ml Congo red (0379; TAGENE); reductive stress-inducing agents: 2 mM DTT (1758-9030; INALCO, United States), 60 μM DMSO (D8418; SIGMA, Germany), 0.6 μM tunicamycin (A611129; BBI, Germany), and 3 μM thapsigargin (A616759; BBI, Germany).

Split-marker knockout vectors were constructed and used for the targeted gene replacement of MoOeIF3K in MoO. To construct split-makers for MoOeIF3K, 0.96 kb upstream and 1 kb downstream flanking fragments were amplified with primers eIF3K-AF/AR and eIF3K-BF/BR, respectively. The upstream fragment was fused into the KpnI and EcoRI enzyme restriction sites at the upstream half of HPH on pCX62. The downstream fragment was fused into the BamHI and XbaI enzyme restriction sites at the downstream half of the HPH on the pCX62 vector by overlap extension PCR cloning (OE-PCR) (Bryksin and Matsumura, 2010). The amplification of PCR products used in the construction of targeted gene replacement vectors was performed using the primer pairs eIF3K-AF + YG/R and HY/F + eIF3K-BR (Supplementary Table 1). The preparation of the MoO protoplast and fungal transformation were performed as described by Talbot et al. (1993) and Nakayashiki et al. (2005). The transformants were screened with eIF3K-OF/eIF3K-OR and eIF3K-UF/eIF3K-UR. The knockout candidates were further confirmed with a Southern blotting assay.

To construct a complementation/MoOeIF3k-GFP fusion vector, a 3,081-bp fragment, including native promoter and whole ORF sequence without stop codon, was amplified. The product was fused into the EcoRI and BamHI enzyme restriction sites upstream of the GFP site on the pKNTG vector. The constructed vector was transformed into the protoplast of the ΔMoOeif3k strain. The transformants were screened with PCR using the primer pairs (eIF3k-OF and GFP-R).

Genomic DNA extraction from the wild-type Guy11, ΔMoOeif3k, and complementation strains was performed with the CTAB DNA extraction procedure described by Brandfass and Karlovsky (2008) and Aliyu et al. (2019). Briefly, the fungal strains were cultured in a CM liquid medium for 4 days at 28°C, 110 rpm. Mycelia were harvested and press-dried with sterilized absorbent filter paper, frozen, and ground into powder in liquid nitrogen. The samples were placed into 2-ml Eppendorf tubes containing 1 ml of cold DNA extraction buffer (100 mM Tris–HCl 8.0, 100 mM EDTA, and 250 mM NaCl) and then vortexed thoroughly; 0.1 ml 10% SDS was added to the mixture and incubated at 37°C for 1 h. Then.15 ml 5 M NaCl and.13 ml 10 M NaCl + 10% CTAB were added to mixture, which was mixed gently and incubated at 65°C for 20 min. After cooling, 0.4 ml chloroform-isoamyl alcohol (24:1 v/v) was added to the mixture, which was thoroughly mixed and centrifuged for 15 min at 10,391 × g. The supernatants were pipetted into new 2-ml Eppendorf tubes containing 2 volumes of chilled 100% ethanol and kept under −20°C overnight to precipitate the DNA. The contents were centrifuged for 15 min at 10,391 × g, and the resulting supernatants were discarded. The precipitated DNAs were air-dried in a laminar air-flow chamber for 5–10 min. The dried DNAs were dissolved with 0.5 ml TE buffer containing 2 μl RNase and incubated for 1 h at 37°C. An equal volume of chloroform:isoamyl alcohol solution (24:1) was later added, mixed thoroughly, and centrifuged for 15 min at 10,391 × g. The supernatants were transferred into new 2-ml EP tubes. A twofold volume of chilled ethanol was added and incubated for about 2 h at −20°C. The contents were centrifuged at 10,391 × g for 15 min, and the resulting supernatants were discarded. The precipitated DNA pellets were washed with 70% ethanol and air-dried. The dried DNA pellets were re-suspended in 100 μl TE buffer and used as templates for PCR amplifications and other experiments.

The strains were cultured in liquid CM for 4 days at a speed of 110 rpm. The mycelia were filtered, washed with sterilized ddH₂O, dried with absorbent paper, and further dry frozen in liquid nitrogen. Total RNA was extracted from the individual strains using a HiPure Universal RNA kit (R4130-02; Magen, China). The expression of MoOeIF3K under different stress conditions and expression of individual subunits of the MoeIF3 complex in ΔMoOeif3k were monitored by quantitative real-time PCR (qRT-PCR) assays. Reverse transcription of RNAs was performed using the PrimeScript RT regent Kit with gDNA Eraser (RR047A; Takara, Japan). A 10-μl reaction mix was formulated as follows: 5 μl TB green, 3.4 μl RNase free water, 0.3 μl of each 10 μM forward and reverse primers listed in Supplementary Table 1 and 1 μl cDNA template. qRT-PCR was carried out with Eppendorf Realplex2 master cycler (AG 223341; Eppendorf, Hamburg, Germany). The raw qRT-PCR data were analyzed using the formula delta delta-CT (2^–ΔΔCT) method described by Rao et al. (2013) and Abdul et al. (2018). The expression level of tubulin was used as the reference or internal control. Error bars represent mean ± SD. The data were obtained from three independent biological experiments with three technical replicates for each independent experiment.

The MoOeIF3k-GFP strain was cultured in a CM liquid medium for 4 days at a speed of 110 rpm. The mycelia were filtered and ground. About equal quantities (weight) of the ground samples were shared to a new CM liquid medium supplemented with different types of stress-inducing agent and MM medium, and incubated for 2 h at a speed of 110 rpm. Filtered individual samples were put into liquid nitrogen after drying out. The samples were ground into fine powder in liquid nitrogen, re-suspended in 5 ml protein lysis buffer (10 mM Tris–HCl, pH7.5, 150 mM NaCl, 0.5 mM EDTA, 0.5% NP40 for 1 L), well-mixed, and kept on ice for 30 min. The contents were centrifuged at 12,074 × g for 15 min at 4°C to remove cell debris. A protein analysis was performed using 12% SDS-PAGE gel and the samples were transferred to polyvinylidene fluoride (PVDF) membranes for Western blot assay. Anti-GFP and anti-actin were purchased from Abmart (United States) and Zoonbio (China), respectively, and used to detect total proteins in cell lysates from the individual strains.

For the Western blotting assays performed to monitor the expression of MoOeIF3k in response to the different types of stress-inducing osmolytes, the MoOeIF3k protein extracted from MoOeIF3k-GFP strains cultured on CM was used as normalization control (NC) for DTT, CR, and MM treatment while the expression level of MoOeIF3k protein recorded for MoOeIF3k-GFP strains pretreated with DMSO was used as the NC for Tu and Tg treatments. Actin-antibody was used as as the Western blot loading control, anti-beta Actin mAb from Zoonbio Biotechnology Co., Ltd., China catalog no. TE0303.

Conidia were washed from 10 day-old culture plates with sterilized ddH₂O and filtered through three layers of lens paper into a 2-mL EP tube. For appressorium formation bioassays, 20 μl spore suspensions (concentration of 5 × 10⁴ spores per ml) were placed as droplets on hydrophobic coverslips and incubated under humid and dark conditions with a temperature of 28°C for 2, 4, 8, 16, and 24 h. More than 100 conidia were examined for each strain per experiment, with consistent results obtained from at least three independent biological experiments with three technical replicates. The turgor pressure level in the appressorium produced by conidia obtained from the individual strains on hydrophobic coverslips were assayed by treating appressorium with 1, 2, 3, and 4 M glycerin solution for 5 min. The rate of appressorium collapse was observed under a light microscope and used to measure appressorium integrity or turgor. At least 100 conidia/appressoria were observed per replicate; in all, a total of three biological experiments, with each consisting of three technical replicates, were carried out in this study.

Conidiophore staining assay was performed by cutting blocks of RBA media fully colonized by the individual strains. The cut blocks were placed on micro-slides with the side bearing the hyphae made to rest on the slide surface and incubated under light for 48 h at 28°C. After this period, the blocks were removed and stained with lactophenol cotton blue (100 ml LCB) solution (LCB solution; 20 ml phenol, 0.6 g cotton blue, 44 ml glycerine, and 16 ml lactic acid; distilled water was added to attain a final volume of 100 ml; the solution was diluted threefold before usage) for 5 min. Finally, the slides were washed with distilled water and visualized under an Olympus Bx51 microscope (Japan) using the bright field mode.

The mating assay was performed with the standard tester strain KA3 (MAT1-1, gifted by Dr. Didier Tharreau) with ΔMoOeif3k and Guy11 (MAT1-2) apart on an oatmeal agar (OA) medium at 20°C for 3–4 weeks.

For hyphae-mediated infection experiments, the individual strains were pre-cultured in liquid CM for 3 days in a shaking incubator at a speed of 120 rpm and a stable temperature of 28°C. The mycelia were filtered out and washed with sterilized double-distilled water. The excess water was drained off. The media-free mycelia were used as propagules to inoculate intact and injured barley leaves. The inoculated plants were first incubated in a dark chamber with stable relative humidity of 90% and a stable temperature of 25°C for 24 h. After 24-hpi, the inoculated tissues were transferred into a growth chamber with a photoperiod of 12-h light/12-h dark.

Under conidia mediated infection assays, spore suspensions were prepared with conidia obtained from the individual strains in a concentration of about 2–5 × 10⁴ per ml fortified with.02% v/v of Tween20 (addition of Tween20 helped in adhering the spores to the leaf tissues). The conidia suspensions were used to spray-inoculate 3 weeks-old blast-susceptible rice seedlings (O. sativa cv. CO39). The inoculated seedlings were incubated under the conditions described above for hyphae-mediated infection, under a dark and humid condition at 25°C for 24 h and transferred into the growth chamber with a 12-h light/12-h dark cycle. In both hyphae- and conidia-mediated infection assays, disease development and lesion severity were assessed 7 days post inoculation (dpi) and used as a measure of pathogenicity and virulence characteristics of the individual strains. Histopathological examinations (host penetration and colonization assays) were performed by inoculating the underside of the barley leaves with conidia suspension with a concentration of 2–5 × 10⁴ per ml. The inoculated tissues were incubated under the set of conditions stated above. The host invasion and colonization efficiencies of the individual strains were observed at 24-hpi by scanning the epidermal tissues of the inoculated barley leaves under a light microscope.

For glycogen staining, the appressoria were stained with a solution prepared from 60 mg/ml of KI and 10 mg/ml of I₂ in distilled water (Thines et al., 2000; Xu et al., 2011; Zheng W. et al., 2015). Yellowish-brown glycogen deposits were visible immediately in the bright field. The samples were stained with Bodipy (D3922; Invitrogen, United States) containing10 μg/ml in phosphate-buffered saline (PBS) (Baerga et al., 2009; Zheng W. et al., 2015) during the comparative evaluation of lipid bodies.

For microscopy, an Olympus DP72 fluorescent microscope or the Nikon A1 Plus confocal microscope (Japan) was used to observe the fluorescent light of GFP and mCherry. The emission and excitation wavelengths were 488 and 561 nm, respectively.

To confirm the localization of MoOeIF3k, we constructed the MoOeIF3k-GFP vector and co-transformed MoOeIF3k-GFP with His-mCherry and Kar2-mCherry markers into Guy11. The transformants were screened by PCR using the primer pair (eIF3K-OF and GFP-R) before using the microscope to screen the transformation strains. The His-mCherry (Zhang et al., 2019) and ER markers (Kar2-mCherry) (Zheng H. et al., 2015) were obtained from Dr. Lianhu Zhang at Jiangxi Agricultural University and Dr. Huawei Zheng at the Minjiang University.

For the immunoprecipitation of GFP fusion proteins from cellular extracts, the total proteins of MoOeIF3k-GFP and GFP strains were extracted and incubated with 30 μl of anti-GFP magarose beads (Smart-Life Sciences, China) for 4 h at 4°C. We then used a magnetic frame to wash the beads three times with 500 μl cold wash buffer (50 mM Tris, 0.15 M NaCl, and pH 7.4) and resuspended the anti-GFP magarose beads in 80 μl SDS-loading buffer. Proteins eluted from the anti-GFP magarose beads were analyzed by immunoblot detection with the anti-GFP antibodies (Abmart, China), followed by mass spectrometry (BGI, China).

Total proteins extracted from the mutant and wild-type strains were quantified using a BCA protein assay kit (BL521A; Biosharp, China). For a BCA working solution, BCA reagents labeled A and B were mixed in the ratio of 50:1.

Next, 20 μl of protein extracted from the individual strains were pipetted into 96-well microtiter plates (optimization was performed according to the instructions of the manufacturer), and 200 μl of the BCA working solution was added to the protein samples, mixed thoroughly with a micropipette, and kept under 37°C for 30 min.

Quantification of total protein content in the samples was recorded by taking readings at an absorbance of 562 nm using a multifunctional microplate reader (LB942; Berthold, China).

To generate clone vectors and positive transgenic yeast strains for Y2H screening of probable MoeIF3k and subunits of the MoOCsn complex, the full-length cDNA of MoOeIF3k was amplified and cloned into a pGBKT7 plasmid containing the GAL4 DBD to obtain the bait vector MoOeIF3k-BD. The full-length cDNAs of MoOCsn1, MoOCsn2, MoOCsn3, MoOCsn4, MoOCsn5, MoOCsn6, MoOCsn7, and MoOCsn12 were amplified and cloned into pGADT7 plasmid harboring the Gal4 activation domain (AD) to obtain the individual prey vectors MoOCsn1-AD, MoOCsn2-AD, MoOCsn3-AD, MoOCsn4-AD, MoOCsn5-AD, MoOCsn6-AD, MoOCsn7-AD, and MoOCsn12-AD according to the protocol described by Young (1998). The interaction of pGBKT7-53 and pGADT7-T was used as the positive control, and pGBKT7-Lam and pGADT7-T were used as negative controls. The resultant bait and prey vectors were confirmed by sequencing and were co-transformed into the yeast strain AH109. All transformants were assayed with 1 × 10⁶ cells/μl droplet on SD-Leu-Trp and SD-Leu-Trp-His-Ade with 20 mg/ml X-α-gal plates.

The yellow fluorescence protein (YFP) was split into two halves fused to MoeIF3k and MoOCsn6 to obtain the MoOeIF3k-NYFP and MoOCsn5-CYFP constructs according to methods described by Zheng et al. (2018). The BiFC constructs were transformed into wild-type protoplasts according to the following combinations: MoOCsn5-CYFP + MoOeIF3k-NYFP, MoOCsn5-CYFP + NYFP, and CYFP + MoOeIF3k-NYFP. The transformants were screened on dual selection culture media containing hygromycin and neomycin. Potential candidates were isolated and confirmed by PCR. A microscopy examination of YFP fluorescence signals in the positive candidates harboring the pair of the split construct was performed with the Nikon A1 laser confocal microscope.