M. oryzae wild-type strain Guy11 and the parental strain “Ku80,” which is wild type Guy11 with a deletion of MoKU80 gene to enhance homologous recombination, were used as control strains throughout this study (Villalba et al., 2008). All strains were stored dried and frozen in cellulose filter paper as described (Valent et al., 1991). For conidiation, strains were cultured on rice bran medium containing 2% rice bran and 1.5% agar (pH 6.0). For testing the utilization of different carbon sources, strains were grown on minimal medium (MM) (Foster et al., 2003) with 10 g/L glucose or an equivalent amount of sucrose, starch or chitin as the sole carbon source. For testing response to chemical stresses, strains were grown on complete medium (CM) (Foster et al., 2003) containing 300 mg/L Congo red, 7.5 mmol/L H₂O₂, 0.01% SDS, or 200 μg/mL Calcofluor white (CFW). For all the plate growth assays, similarly-sized plugs of each strain were transferred from starch-yeast medium (SYM) to the appropriate media plates, and plates were imaged after cultivation for 10 days. All assays in this study were performed in three independent biological experiments with at least three replicates. The significance analysis was performed using the Duncan's multiple range test in SPSS software (Bryman and Cramer, 2012). This method was applied in all the other statistic analyses except for the width/length ratio of IH in Figure 5 and the optical density/g mycelia in Figure 7C.

To generate ΔMosec4 mutants, a 616-bp upstream fragment (A) and a 741-bp downstream fragment (B) of the targeted ORF were amplified from the M. oryzae genome using the primer sets MoSEC4-2F and MoSEC4-2R, and MoSEC4-3F and MoSEC4-3R, respectively (Figure S1 and Table S3). The purified products were inserted into the Hind III—EcoR I and BamH I—Spe I sites of plasmid pCX62 that contained the hygromycin B phosphotransferase cassette (H). The “A-H-B” fusion was amplified from pG06135KO (Table S2) using primers MoSEC4-2F and MoSEC4-3R (Table S3) and transformed into the protoplasts of strain Ku80 through PEG-mediated transformation as described previously (Sweigard et al., 1992, 1998). To identify the targeted gene deletion mutants, all hygromycin-resistant transformants were first screened by PCR using primers MoSEC4-4F and MoSEC4-4R, and MoSEC4-5F and MoSEC4-5R (Table S3). The positive transformants were confirmed through Southern hybridization and RT-PCR.

For complementation of ΔMosec4 mutant, a 2.2-kb genomic DNA fragment containing the native promoter, CDS region and 3′-UTR of MoSEC4, was amplified using primers MoSEC4-6F and MoSEC4-6R (Table S3), and cloned into pKNTG plasmid. The resulting construct was reintroduced into ΔMosec4 mutant through PEG-mediated transformation. The neomycin-resistant transformants were evaluated through RT-PCR and Southern hybridization.

To produce the P_MoSEC4:GFP:MoSEC4 transformants for complementation and fusion protein localization, the pBV1171 (Table S2) was introduced into the ΔMosec4 mutant and the transformants were evaluated through PCR-based genotyping using GFPXbaIF-MoSEC4tR and MoSEC4ptF-MoSEC4tR primer sets (Table S3).

5–10 μg genomic DNA of each strain was completely digested with EcoRI. Digested products were electrophoresed in a 1% (m/v) agarose gel and then transferred onto a Hybond-N⁺ membrane (Amersham Pharmacia Biotech Inc.). The specific hybridization probe was amplified using primers MoSEC4-4F and MoSEC-4R (Table S3). Probe labeling, hybridization and detection were performed using DIG High Prime DNA Labeling and Detection Starter Kit I (Roche Applied Science), according to the instruction manual.

After the removal of aerial mycelia, culture blocks (4 mm²) were cut from 5-day-old rice bran agar cultures, laid on their sides on slides and placed into moisturizing plastic plates. After the induction of conidiogenesis for 24 h in a chamber with constant-light conditions, conidial development was observed using an Olympus BX53 microscope (Liu et al., 2010).

To analyze the conidial morphology, conidia of each strain collected from 10-day-old rice bran agar cultures were quantified using a hemocytometer, and were observed for morphology under light microscopy. The length and width of spores and IH cells were measured using the line tool of ZEN 2010 software (Zeiss).

To examine conidial germination and appressorial formation, 10 μL of conidial suspension (5 × 10⁴ conidia/mL) were dropped on the hydrophobic side of Gelbond film (Lonza), and then incubated in a moist chamber at 28°C. The percentages of conidial germination and appressorial formation were determined by microscopic examination at 4, 8, 12 hpi.

The cytorrhysis assay was performed for measuring appressorium turgor as described (de Jong et al., 1997). In brief, conidia were incubated on the surface of artificial hydrophobic Gelbond film, and treated with different concentrations of glycerol (1M, 2M, 3M) after appressorium formation (at 24 h post germination). Collapsed, sometimes also ruptured, appressoria were counted using DIC microscopy 5 min after glycerol treatment.

Conidial suspensions (1 × 10⁵ conidia/mL in 0.25% gelatin solution) of each strain were used for inoculation assays on 15-day-old rice seedlings of susceptible rice varieties CO39 and YT16. Rice seedlings were cultivated under the conditions described previously (Giraldo et al., 2013). Plant inoculation and subsequent incubation were performed as previously described, with the humidity maintained below 90% to block conidation (Valent et al., 1991). Lesion formation was examined at 7 days post inoculation. Lesion types were defined according to Valent et al. (1991). Specifically, Type 0 indicated no visible evidence of infection; type 1, uniform dark brown pinpoint lesions up to 0.5 mm long without visible centers; type 2, small lesions with distinct tan/green centers surrounded by a darker brown margin (up to 1 mm); type 3, small eyespot lesions ~1.5 mm in length with tan centers surrounded by dark brown margins; type 4, intermediate size eyespot lesions, up to 2 mm in length; type 5, large eyespot lesions that attain the maximum size characteristic for a particular cultivar/strain interaction (~3 mm in length for Guy11 on C039). Lesion numbers were counted using 5 cm-long diseased rice blade sections taken from the youngest leaves at the time of inoculation. Assays were repeated at FAFU and KSU with similar results.

Leaf sheath inoculation was performed using 4–5 week old rice plants as described (Kankanala et al., 2007). For the penetration rate assay, the spore solutions (5 × 10⁴ conidia/mL) were inoculated into the leaf sheath of susceptible rice line YT16 as described (Giraldo et al., 2013) and examined at 28 h after inoculation.

A 3 × 3 mm hyphal tip plug was inoculated into fresh CM liquid media and incubated at 28°C, 150 rpm for 3 days. Mycelia were completely removed by filtration, and the culture filtrates were used for the measurement of extracellular enzyme activities. Laccase activity was measured using the ABTS method as described previously (Song et al., 2010). The activities of amylase, invertase, and cellulase were determined using the DNS method as described (Miller, 1959). The dry weights of the harvested mycelia were measured for normalizing the enzyme activities.

The live-cell imaging assay was performed as described previously (Giraldo et al., 2013). For vegetative hyphal imaging, a small plug of mycelium was cut from an agar culture, placed on the edge of a sterile water agar-coated microscope slide and incubated in a humid chamber for 16–18 h. For imaging invasive hyphae, leaf sheath inoculation was performed using the susceptible rice line YT16. Differential interference contrast microscopy was performed with a Zeiss Axioplan 2 IE MOT microscope, using × 40/0.75 and × 63/1.2 NA (numerical aperture) C-Apochromat water immersion objective lens. Images were acquired using a Zeiss AxioCam HRc camera and analyzed with Zeiss Axiovision digital image-processing software (version 4.8). Confocal imaging was performed with a Zeiss LSM780 confocal microscope system using two water immersion objectives, C-Apochromat 40x/1.2 W Corr and C-Apochromat 40x/1.2 W Corr. Images were acquired and processed using ZEN 2010 software.

To identify the Sec4 ortholog in M. oryzae, the amino acid sequence of Sec4p from the S. cerevisiae genome database (http://www.yeastgenome.org/) was used for a BLAST search in the M. oryzae genome database (http://www.broadinstitute.org/ annotation/genome/magnaporthe_grisea/MultiHome.html). This search identified MGG_ 06135.7, termed MoSec4, as the ortholog of S. cerevisiae Sec4p. The MoSec4 protein shares high amino acid identity with its orthologs, such as CoSec4 in Colletotrichum orbiculare (BAO27795.1, 89% identity), BcSas1 in Botrytis cinerea (BC1G_14039.1, 87% identity), Sec4p in S. cerevisiae (NP_116650.1, 64% identity), and both Rab8A (NP_005361.2, 65% identity) and Rab8B (NP_057614.1, 63% identity) in Homo sapiens. Multiple sequence alignment indicated that all the above six proteins possess the signature motifs of Rab GTPases, including highly conserved Rab-specific residues RabF1-F5, Rab subfamily-specific sequences RabSF1-SF4, highly conserved motifs involved in guanine and phosphate/Mg²⁺ binding (G1-G3 and PM1-PM3), and a cysteine motif at carboxyl terminus required for subcellular localization (Figure 1).

To study the function of MoSec4, we generated targeted gene replacement mutants of the MoSEC4 gene in an M. oryzae parental strain Guy11_ΔMoKu80 (hereinafter referred to as “Ku80”), which is wild type Guy11 with a deletion of the MoKU80 gene to enhance homologous recombination (Villalba et al., 2008). Transformants were produced in which the ORF of the MoSEC4 gene was replaced with a hygromycin B phosphotransferase (HPH) gene (Figure S1A). Two MoSEC4 deletion mutants, ΔMosec4-11 and ΔMosec4-46, were characterized by reverse transcription polymerase chain reaction (RT-PCR) and Southern blot analysis (Figures S1B,C). To confirm phenotypes associated with mutation of the MoSEC4 gene, the entire ORF with its native promoter was introduced into the ΔMosec4-11 mutant (hereafter known as ΔMosec4), and a complementation transformant ΔMosec4-Com was also confirmed by Southern analysis (Figures S1B,C).

To explore the role of MoSec4 in vegetative growth, the parental strain (Ku80), ΔMosec4 mutants and complemented mutant (ΔMosec4-Com) were grown on SYM, RDC, and CM media. The ΔMosec4 mutants exhibited a significant reduction of growth rate on all the media (Figure 2A and Figure S2). In addition, the ΔMosec4 mutants displayed more branching and were prone to form swollen hyphal tips (Figure S3). Additionally, the fluorescence of the polarized growth marker MoExo70-GFP fusion was not observed at mutant hyphal tips (Figure S4). This suggested that filamentous hyphal tips frequently undergo depolarization in the mutants.

We also evaluated the role of MoSec4 in conidiation of M. oryzae. The results showed that the ΔMosec4 mutants produced significantly fewer conidia compared with the parental (Ku80) and ΔMosec4-Com strains (Figure 2B). Consistent with this result, we observed that the number of conidia per conidiophore was reduced in ΔMosec4 mutants (Figure 2C). This result indicated that deletion of MoSEC4 caused a defect in sympodial conidiogenesis and subsequent reduction of conidium production. Furthermore, conidia from the ΔMosec4 mutants displayed abnormal morphology. Specifically, mutant conidia lost the pyriform-shape and showed a higher length-width ratio than those of the parental and complemented strains (Figure 2D and Table S1). Taken together, these results suggested that MoSec4 plays a role in conidiogenesis of M. oryzae.

In S. cerevisiae, deletion of Sec4p leads to the abnormal deposition of chitin, which is an important component of fungal cell walls. To determine if MoSec4 is involved in response to stresses reported to be associated with cell wall integrity, we investigated the growth of ΔMosec4 mutants on media containing each of several stress-inducing chemicals, including Congo Red, H₂O₂, SDS, and CFW. Among them, Congo Red binds to glucan and acts as a cell wall perturbing agent (Wood and Fulcher, 1983; Song et al., 2010). SDS is a detergent used to test cell wall integrity, because a defect in cell wall integrity will increase the accessibility of SDS to damage the plasma membrane (Shimizu et al., 1994; Igual et al., 1996; Bickle et al., 1998). CFW binds to nascent chitin and glucan (with less affinity), perturbs the microfibril assembly and subsequently impairs the cell wall integrity (Elorza et al., 1983; Ram et al., 1994; Lussier et al., 1997). Additionally, H₂O₂ causes oxidative stress (Bai et al., 2003). Growth of ΔMosec4 mutants on media containing Congo Red, H₂O₂, SDS, and CFW was severely inhibited, especially on media with CFW (Figure 3). These results suggested that MoSec4 may be involved in response to different stresses, including stresses associated with integrity of hyphal cell walls.

To investigate whether MoSec4 was involved in pathogenicity, conidial suspensions of Ku80, ΔMosec4, and ΔMosec4-Com strains were inoculated respectively, onto 3-week-old rice seedlings of susceptible rice cultivar CO39. At 7 days post inoculation (dpi), the parental strain and complemented transformant caused Type 3 to Type 4 lesions with green/tan centers that are capable of abundant sporulation (Valent et al., 1991). In contrast, the ΔMosec4 mutants produced Type 1 hypersensitive flecks that fail to sporulate or Type 2 lesions with small green/tan centers and limited sporulation potential (Figure 4A and Figure S5A). Moreover, the number of lesions produced by the ΔMosec4 mutants was decreased (~50%) relative to the parental strain and complemented mutant (Figure S5B).

The attenuated pathogenicity of ΔMosec4 mutants could be caused by a defect in growth, infection-related development and/or the secretion of effectors. In order to understand what caused the defect, we first assessed the effects of MoSEC4 deletion on germination of conidia as well as formation and maturation of appressoria. Deletion of MoSEC4 did not affect the conidial germination, and it retarded, but did not block appressorial formation (Figure S6). Therefore, neither process appears responsible for the observed pathogenicity defect.

We next examined the turgor pressure in appressoria of parental strain Ku80 and ΔMosec4 mutants using a standard cytorrhysis assay, in which appressoria collapse when the external glycerol concentration exceeds the concentration of intracellular osmolytes, such as glycerol (de Jong et al., 1997). The results showed that, after treatment with 2 M glycerol for 5 min, only 25% of appressoria produced by the parental or complemented strains collapsed, whereas, under the same conditions, more than 60% of appressoria produced by the ΔMosec4 mutant collapsed (Figures 4B,C). Therefore, appressoria produced by the mutant had lower levels of turgor pressure, which could account for lower levels of host cuticle penetration and lower numbers of lesions in leaf infection assays.

We also performed a leaf sheath inoculation assay to document the appressorial penetration rates and biotrophic invasion of Guy11, Ku80, and ΔMosec4 mutants on susceptible rice YT16. Consistently, the penetration rate of the ΔMosec4 mutant at 28 hpi was ~25% compared with Guy11, which is significantly reduced compared to penetration rates of ~70% for Ku80 compared to Guy11 (Figure 4D). Penetration by the mutant strain was ~36% of the rate of penetration by the parental strain Ku80. Live-cell imaging was performed to observe the morphology of IH produced by the ΔMosec4 mutant inside rice cells. The results showed that the individual bulbous IH cells of the ΔMosec4 mutant were often shorter and swollen, which makes the mutant IH appear compressed (Figure 5A). For the ΔMosec4 mutant, 91% of the IH (110 of 121 randomly imaged IH) had shorter and swollen bulbous IH cells, compared to 2.7% of those for Ku80 (2 of 74 randomly imaged IH). A detailed analysis of the length to width ratio of the subapical bulbous IH cells (location indicated in Figure 5A) showed that the average length to width ratio of tested cells is 1.56 for ΔMosec4 mutant IH cells, which is significantly less than the ratio of 2.77 for Ku80 (Figure 5B). Thus, there is a morphological effect on mutant IH, which could potentially impact leaf colonization and lesion development.

To investigate the localization of MoSec4 protein, the MoSEC4 gene was expressed with its native promoter and an N-terminal GFP fusion (MoSEC4p:GFP:MoSEC4) and introduced into the Ku80 and ΔMosec4 strains (Figure S7A). The phenotype of ΔMosec4_GFP-MoSEC4 transformants was analyzed to assess whether the GFP-MoSec4 fusion protein was functional. Our microscopy results showed that the expression of GFP-MoSec4 fusion could restore wild type spore shape (Figure S7B and Table S1). The GFP-MoSec4 fusion protein also complemented the attenuated pathogenicity of ΔMosec4 mutant (Figure 6A). Genotyping confirmed the insertion of the P_MoSEC4:GFP:MoSEC4 fusion in all the transformants (Figure S7C). Taken together, these results suggested that the GFP-MoSec4 fusion protein is functional and again confirms that defects described in the mutant are caused by deletion of MoSEC4 gene.

Live-cell imaging of growing vegetative hyphae was performed for localization of the GFP-MoSec4 fusion protein. The fluorescent MoSec4 protein mainly localized in crescents adjacent to the tips of growing hyphae (Figure 6B), which is consistent with a conserved function in secretion of hyphal cell wall components and apical growth.

To determine whether MoSec4 is required for the secretion of extracellular enzymes, we cultured both the Ku80 and the ΔMosec4 mutant on MM media containing different disaccharides or polysaccharides as the sole carbon source to examine the secretion of relevant hydrolases. We found that the mycelial growth of ΔMosec4 mutants was dramatically inhibited (Figure 7A). When grown on the MM media containing sucrose, starch, or chitin, the mycelial growth rate of ΔMosec4 mutants was inhibited by 36, 48, and 46%, respectively, compared to growth on glucose (Figure 7B). These results suggested that MoSEC4 deletion partially blocked the secretion of several extracellular hydrolases, invertase, amylolytic enzyme and chitinase (respectively involved in the enzymolysis of sucrose, starch, and chitin) and subsequently inhibited the utilization of these carbon sources. Furthermore, we directly determined the activity of carboxymethyl cellulase (CMCase), fructosidase and amylase in fungal culture filtrates, and the results showed that the activity of these extracellular enzymes was dramatically decreased for the ΔMosec4 mutant (Figure 7C). Taken together, our data suggest that MoSec4 plays an important role in the secretion of extracellular enzymes.

To assess whether the MoSec4 protein is also involved in the secretion of effectors, we performed live cell imaging of biotrophic invasion by mutant strains expressing cytoplasmic effector fusion protein PWL2-mCherry-NLS (with an added nuclear localization signal “NLS” to facilitate detection of translocated effector protein in rice cells), and apoplastic effector fusion BAS4-GFP (Khang et al., 2010). As expected, in the Guy11 and Ku80 strains, the BAS4-GFP fusion protein outlines the IH and occupies an inner layer of the BIC, and the PWL2-mCherry-NLS protein mainly concentrated in BICs and was translocated into rice nuclei. In contrast, the localization of PWL2-mCherry-NLS for the ΔMosec4 mutant was partially disrupted, while localization of the BAS4-GFP fusion appeared normal (Figure 8). Localization patterns of PWL2-mCherry-NLS in ΔMosec4 mutant could be further divided into three groups as shown in Figure S8. Among all the observations, 34.1% of the randomly imaged infection sites (30 in 88) were normal (type I), 39.8% (35 in 88) of them formed additional small fluorescent punctae outside of BICs (type II), and 26.1% (23 in 88) of them formed two or more BIC-like structures (type III). In contrast, only 15.6% (5 in 32) of the randomly imaged infection sites in Ku80 showed the type II pattern, and none showed type III. Most of the Ku80 sites appeared normal. Therefore, proper localization of cytoplasmic effectors in BICs depends in part on MoSec4.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.