Magnaporthe oryzae fimbrin organizes actin networks in the hyphal tip during polar growth and pathogenesis

Magnaporthe oryzae causes rice blast disease, but little is known about the dynamic restructuring of the actin cytoskeleton during its polarized tip growth and pathogenesis. Here, we used super-resolution live-cell imaging to investigate the dynamic organization of the actin cytoskeleton in M. oryzae during hyphal tip growth and pathogenesis. We observed a dense actin network at the apical region of the hyphae and actin filaments originating from the Spitzenkörper (Spk, the organizing center for hyphal growth and development) that formed branched actin bundles radiating to the cell membrane. The actin cross-linking protein Fimbrin (MoFim1) helps organize this actin distribution. MoFim1 localizes to the actin at the subapical collar, the actin bundles, and actin at the Spk. Knockout of MoFim1 resulted in impaired Spk maintenance and reduced actin bundle formation, preventing polar growth, vesicle transport, and the expansion of hyphae in plant cells. Finally, transgenic rice (Oryza sativa) expressing RNA hairpins targeting MoFim1 exhibited improved resistance to M. oryzae infection, indicating that MoFim1 represents an excellent candidate for M. oryzae control. These results reveal the dynamics of actin assembly in M. oryzae during hyphal tip development and pathogenesis, and they suggest a mechanism in which MoFim1 organizes such actin networks.


Introduction
Magnaporthe oryzae is a filamentous fungus and the causal agent of rice blast disease. Each year, infections caused by this recalcitrant pathogen destroy up to 30% of the rice (Oryza sativa) crop and threaten global food security [1,2]. M. oryzae undergoes extensive developmental changes while building elaborate infection structures, invading plant cells, and finally proliferating inside the host cells [3,4]. Thus, understanding the biology of M. oryzae development and infection is critical for developing durable strategies for controlling rice blast disease.
In filamentous fungi, hyphae generally extend by tip growth [5,6]. This process is thought to be associated with the polarized trafficking of secretory vesicles to the Spitzenkörper (Spk), the organizing center for hyphal growth and development [7]. Cell tip growth determination and protein secretion are intriguing issues in fungal cell biology [6]. Polarized exocytosis is an essential process in fungi and is involved in cell morphogenesis and pathogenesis [8,9]. The Spk, which acts as a vesicle supply center, is composed of a dense group of vesicles, cytoskeleton, ribosomes, and other undefined components delivered in a polar manner [7,10,11]. The exocyst complex in M. oryzae is located at the hyphal tip, ahead of the Spk, indicating that a successive protein secretion pathway is required for fungal infection [12]. However, even though the Spk plays critical roles in the development and pathogenesis of filamentous fungi, how the Spk is maintained and moves in a polar manner in fungal cells is not well understood.
Actin is a highly conserved component of the cytoskeleton found in all eukaryotes and plays important roles in a variety of cellular processes [13,14]. The main components of the actin cytoskeleton include monomeric globular actin (G-actin), polymeric filamentous actin (F-actin), motors, actin-binding proteins, and septins (which serve as actin scaffolds) [15]. During the past decades, much progress has been made in elucidating the crucial roles of the actin cytoskeleton in fungal morphogenesis and the pathogenesis of M. oryzae. For example, the M. oryzae actin motor proteins MoMyo2 and MoMyo5 are associated with hyphal growth and appressorial development; thus, their mutants exhibit defects in the infection of rice host cells [16][17][18]. The actin-binding proteins MoCAP and MoEnd3 are key factors in M. oryzae development and virulence through regulating actin assembly [19,20]. M. oryzae septins polymerize into a heteromeric ring to scaffold a toroidal F-actin network at the base of the appressorium to facilitate the formation of a rigid penetration peg, which breaches the leaf surface [21]. Pmk1, a fungal mitogen-activated protein kinase (MAPK), was recently found to regulate the expression of the actin cross-linking protein alpha-actinin, and the septin phosphorylation protein Chm1, thus controlling the hyphal constriction required for the fungus to penetrate from one rice cell into its neighbor [22]. These findings shed light on the importance of the actin cytoskeleton in M. oryzae development and pathogenesis. However, despite this progress, live-cell visualization is needed to explore the dynamic remodeling of the actin cytoskeleton in M. oryzae during typical cellular processes, including development and infection, to understand its versatile functions.
Fimbrin is a conserved F-actin cross-linking protein present in organisms ranging from yeast to mammals and plays important roles in numerous fundamental cellular processes.
Fission yeast fimbrin (Sac6) is reported to play important roles in endocytosis, cytokinesis, and polarization [23,24]. Five FIMBRIN genes are present in the Arabidopsis thaliana genome [25,26]. Among the encoded fimbrins, FIMBRIN1 and FIMBRIN5 have been implicated in the regulation of pollen development and FIMBRIN5 might regulate the generation and maintenance of the rigidity of actin bundles oriented along the length of the pollen tube [26][27][28]. In addition, GFP-ABD2, comprising the second actin-binding domain (ABD2) of Arabidopsis FIMBRIN1 fused with green fluorescent protein (GFP), is widely used as an actin-labeling probe in plants [29][30][31].
Compared to our understanding of fimbrin functions in plants, our knowledge of the physiological functions of fimbrin in filamentous fungi is limited. A fimbrin protein in Aspergillus nidulans was observed as mobile patches throughout the hyphae and concentrated near hyphal apices, which are thought to play a role in endocytosis [32]. A M. oryzae fimbrin protein was previously reported to accumulate in dynamically moving actin patches at the hyphal subapical region and function by interacting with the exocyst [12]. However, neither the mechanisms by which filamentous fungal fimbrins regulate the dynamic organization of actin filaments, nor the associated underlying cellular processes, are well understood. Addressing these issues would help unravel exactly how fimbrin regulates the growth and pathogenesis of fungal pathogens, increasing our understanding of the functions of the actin cytoskeleton in filamentous fungi.
In this study, we labeled the actin cytoskeleton in the model pathogenic fungus M. oryzae using the actin probe Lifeact-GFP [33,34] and performed super-resolution live-cell imaging to reveal the architecture and dynamics of the actin cytoskeleton during polar growth and vesicle delivery in fungal hyphae. We demonstrate that the actin cytoskeleton forms a unique structure at the apex of the active hyphae, including actin at the Spk and the formation of branched actin cables originating from the Spk and extending into the cell membrane. Furthermore, we show that M. oryzae Fimbrin (MoFim1) helps maintain this elaborate actin organization. Our results suggest that the development and elaboration of apical actin structures help ensure vesicle delivery for fungal development and protein secretion as well as hyphal expansion in rice cells. In addition, this study increases our understanding of the actin-based regulatory mechanism underlying both the development and pathogenesis of filamentous fungi.

MoFim1 is important for the polar growth, conidiation, and full virulence of M. oryzae
The physiological roles of fimbrins in regulating actin rearrangement during M. oryzae development and pathogenesis have not been elucidated. The M. oryzae Y34 genome contains one Fimbrin gene, MoFim1. To investigate the physiological functions of MoFim1, we employed the standard one-step gene replacement strategy to knock out MoFim1 in M. oryzae [35] (S1 Fig). When cultured on complete medium (CM) or straw rice bran (SRB) medium, the Mofim1 knockout mutant developed at a significantly lower rate compared to the wild type (WT) and complemented strain (Fig 1A and 1B, S2 Fig). Furthermore, conidiation was completely abolished in the Mofim1 mutant. No mature spores were present on the conidiophores, as observed by microscopy ( Fig 1C).
Because the mutant fails to produce mature spores, we could not carry out single-spore isolation. Instead, we generated protoplasts using three Mofim1 deletion clones. We diluted and cultured the protoplasts on CM agar medium, then picked and cultured a single colony from each clone. We found that all colonies exhibited the same developmental phenotype as the Mofim1 mutant. Furthermore, all the single colonies could be recovered to a wild-type

PLOS PATHOGENS
Fimbrin organizes actin assembly in M. oryzae phenotype by expressing MoFim1 driven by its native promoter (S3A Fig). We also performed a Southern blot experiment to examine these mutants, which showed that these colonies were true deletion mutants (S3B Fig). We next analyzed the pathogenesis of Mofim1 using the same region on the plate for SRBgrown cultures of WT, Mofim1, and the complemented strain to infect both abraded and nonabraded rice leaves (Oryza sativa cv. Nipponbare). We found that deletion of MoFim1 severely affected the ability of M. oryzae to infect rice leaves. Comparing with the WT and complemented strains, the Mofim1 mutant induced a rather small lesion on the abraded rice leaves (Fig 1D and 1F). On the non-abraded leaves, we could not find evidence for penetration and infection by the mutant (Fig 1E and 1F).
Microscopy of the hyphae revealed that while almost all WT and complemented tips maintained the hyphoid shape, the Mofim1 mutant had more blunt ends (Fig 1G). These results indicate that MoFim1 plays important roles in the development and pathogenesis of M. oryzae.

MoFim1 helps organize the actin cytoskeleton in the hyphal tip
As fimbrin is a putative actin-binding protein, we labeled the actin cytoskeletons of the WT and Mofim1 hyphae with the widely used Lifeact-GFP peptide probe [21,33,34]. Given that the organization of actin in developing M. oryzae hyphae is not well understood, we examined the dynamic distribution of actin filaments during M. oryzae development using a super-resolution live-cell imaging system. In growing WT vegetative hyphae, a population of actin filaments was continuously generated from the Spk area; the generation of these filaments was tightly linked with hyphal growth (Fig 2A and S1 Movie). In addition to the actin at the Spk and in actin cables, we also observed actin accumulation in patches at the subapical collar region (Fig 2B and 2C). Images from Z-slice projections show that the subapical actin patches

PLOS PATHOGENS
Fimbrin organizes actin assembly in M. oryzae mainly clustered close to the cell membrane ( Fig 2C, slice Z-1 and Z-2). The actin cables that originated from the Spk formed long, branched actin bundles radiating into the subapical region; these bundles appeared as a bowl-shaped structure at the hyphal head ( Fig 2D and S2 Movie). Furthermore, we noticed that some patches travelled dynamically along the actin cables ( Fig 2E and S3 Movie). We developed a schematic representation of actin organization and remodeling associated with hyphal growth based on our observations ( Fig 2F). These results uncover a unique actin organization pattern underlying elongation that occurs via tip or diffuse growth.
The hyphae of WT and complemented lines produced a population of dense actin filaments converging in the Spk during growth (Fig 3A and 3C, S4A and S4C Movie). The mean growth rate was~0.79 and 0.80 μm/min for the WT and the complemented hyphae, respectively ( Fig  3D, S5A and S5C Movie). However, in Mofim1 hyphae, we observed intermittent appearance of actin at the Spk and the average growth rate was only~0.25 μm/min (Fig 3D, S5A and S5B Movie). When the mutant hyphae were growing, the hyphal tips gradually became pointed and, at the same time, actin was observed at the Spk and Spk-converging actin filaments formed (from 3:45 [min:s] to 6:00 in Fig 3B). The growth rate in this time period was 0.68 μm/min. Subsequently, the actin at the Spk gradually disappeared and the actin filaments became loosely distributed in the apical regions of the hyphae. Finally, the growth rate decreased to~0.1 μm/min and the hyphal tips began to expand (from 6:00-7:30 in Fig 3B and  S4B Movie). Quantitative analysis showed that actin at the Spk was maintained in mutant hyphae for~30% of the time that it was maintained in the WT and the complemented hyphae during growth (Fig 3E). The area of the actin at the Spk, and the width and number of the actin filaments were significantly reduced in the Mofim1 mutant (Fig 3F-3H). We performed fluorescence recovery after photo bleaching (FRAP) analysis to assess the turnover of the actin cytoskeleton in the WT versus the Mofim1 mutant. Actin fluorescence recovered to a higher level in WT cells compared to Mofim1 cells, indicating that the dynamic state of actin was reduced in the mutant cells ( S4 Fig and S6 Movie). Together, these results indicate that MoFim1 plays important roles in organizing the actin cytoskeleton in the hyphal tip and is therefore crucial for hyphal morphogenesis and tip growth.

MoFim1 colocalizes with the actin cytoskeleton
To investigate how MoFim1 regulates the dynamics of the actin cytoskeleton in the hyphal tip, we observed the localization of MoFim1-mCherry in M. oryzae. MoFim1-mCherry driven by its native promoter and Lifeact-GFP were coexpressed in the Mofim1 mutant, and the colocalization of GFP and mCherry was observed by super-resolution live-cell imaging microscopy. Time-lapse imaging of vegetative hyphae revealed that MoFim1 formed dense patches in the cytoplasm or the actin collar at the subapical region. MoFim1-mCherry also colocalized with the actin at the Spk in rapidly growing hyphae. Some MoFim1-mCherry patches in the cytoplasm localized to actin cables ( Fig 4A and S7 Movie). MoFim1 exhibited a similar distribution in invasive hyphae: MoFim1 accumulated at the cell tip, with strong signals from the Lifeact-GFP-labeled actin cytoskeleton (Fig 4B, S5 Fig and S8 Movie). Line-scan analysis further supported the colocalization of MoFim1-mCherry and the actin cytoskeleton at the Spk and subapical collar (Fig 4C and 4D). Based on our observations, we developed a model of the localization pattern of MoFim1 in hyphae ( Fig 4E). MoFim1-mCherry patches were also visible at the tip of the germ tube, the actin at the periphery of the cell in the appressorium, and the septum (Fig 4F-4H). These results indicate that MoFim1 closely associates with the actin cytoskeleton during polarized hyphal growth.

MoFim1 binds to and bundles actin filaments
As fimbrin is a major regulator of actin organization [26,28], we investigated the biochemical basis for the function of MoFim1. After producing His-tagged recombinant MoFim1 in Escherichia coli, we performed a high-speed co-sedimentation assay to assess the ability of MoFim1 to bind to actin filaments. As shown in S6A Fig, MoFim1 bound to and coprecipitated with actin filaments. When we incubated F-actin (4 μM) with MoFim1, the amount of MoFim1 in the pellet increased in proportion to the MoFim1 concentration (0-8 μM).
Since fimbrin family members generally bundle actin filaments, we performed a low-speed cosedimentation assay to determine whether MoFim1 also possesses this property. As shown in S6B Fig, the amount of F-actin in the supernatant decreased in the presence of increasing levels of MoFim1, indicating that more F-actin bundles formed and precipitated in the pellets.
MoFim1 is composed of an N-terminal EF-hand (EF) motif and two actin-binding domains (ABD1 and ABD2) in tandem ( Fig 5A). To elucidate the actin binding and bundling mechanism of MoFim1, we used an in vitro system to produce GFP-fused full-length and truncated versions of MoFim1, as shown in Fig 5A. We incubated these proteins with polymerized Factin to evaluate their F-actin binding/bundling activities. Single-molecule imaging of phalloidin-stained F-actin revealed fine actin filaments when incubated with the GFP control ( Fig  5B). However, when incubated with MoFim1-GFP, thick actin bundles appeared, and punctate MoFim1-GFP signals were visible on the actin filaments ( Fig 5C). These results indicate that MoFim1 binds to and cross-links F-actin to form actin bundles. Large actin aggregates formed in the presence of Ca 2+ , implying that the actin bundling activity is accelerated by Ca 2+ ( Fig  5D).
We also analyzed the biochemical properties of the MoFim1 domains shown in Fig 5A. The ABD1 and ABD2 domains, but not the EF domain, bound to F-actin. All three domains lacked F-actin bundling activity (Fig 5E-5G). Moreover, when the EF domain was fused to ABD1 and ABD2, both EF-ABD1 and EF-ABD2 bound to and bundled actin filaments (Fig 5H and 5I). These results suggest that the EF domain facilitates the actin binding/bundling activities of MoFim1 via the formation of homodimers. This notion was further confirmed by a yeast twohybrid assay, as shown in S7 Fig. In addition to these in vitro experiments, we performed genetic complementation tests using genomic constructs with various domain truncations to unravel the molecular and cellular basis of MoFim1 activity. None of the truncated versions of MoFim1, including EF, ABD1, In contrast to the distribution of the intact MoFim1 in the Spk, actin cables, and the subapical collar (Fig 6A), EF-mCherry was distributed uniformly inside WT and Mofim1 hyphae ( Fig 6B). ABD1-mCherry accumulated at the hyphal apex and colocalized with actin at the Spk ( Fig 6C). When EF was fused to ABD1, brighter mCherry signals were observed at the Spk. However, this fusion protein was not obviously colocalized with Factin in the subapical region ( Fig 6D). By contrast, ABD2-mCherry labeled both actin at the Spk and F-actin at the subapical region of the M. oryzae hyphae (Fig 6E), and the EF motif promoted its actin labeling efficiency ( Fig 6F).

PLOS PATHOGENS
Fimbrin organizes actin assembly in M. oryzae

MoFim1 is involved in actin cytoskeleton-mediated endocytosis
We then investigated the physiological functions of MoFim1. The actin collar plays a role in endocytosis [7] and some MoFim1 localized to the actin collar, indicating that MoFim1 functions in endocytosis (Fig 4). To verify this notion, we stained the hyphae with FM4-64, a widely used marker of endocytosis. Approximately 5 min after incubation with FM4-64, the red signals were endocytosed in the cytoplasm of the hyphae and delivered to the hyphal tip, where they then colocalized with the actin at the Spk (Fig 7A).
To investigate the roles of the actin cytoskeleton in endocytosis, we treated the hyphae with the actin-depolymerizing drug latrunculin B (Lat B). The actin cytoskeleton, including the actin collar, actin cables, and Spk, were destroyed upon Lat B treatment (Fig 7B). These hyphae failed to undergo polar growth; instead, the hyphal tip began to inflate. Strong red signals from FM4-64 accumulated at the hyphae cell membrane, indicating that the internalization of this endocytotic marker had been delayed. Moreover, the endocytosed FM4-64 was not

PLOS PATHOGENS
Fimbrin organizes actin assembly in M. oryzae transported to the Spk region ( Fig 7B). We next stained Mofim1 hyphae with FM4-64 and found that FM4-64 could be transported to the Spk when actin was accumulating at the Spk ( Fig 7D). Actin appeared only intermittently at the Spk in the Mofim1 mutant ( Fig 3B); when actin at the Spk depolymerized, accumulation of FM4-64 at the Spk was subsequently attenuated ( Fig 7E). Together, these results indicate that MoFim1 is required for actin cytoskeletonmediated endocytosis.

MoFim1 organizes the actin cytoskeleton in the hyphal tip for septindependent assembly of the exocyst
The failure of Lat B-treated WT and Mofim1 hyphae to accumulate of FM4-64 at the Spk ( Fig  7B and 7E) indicated that the delivery of secretory vesicles is dependent on the actin cytoskeleton in the hyphal tip. The septin-dependent exocyst in M. oryzae is located near the Spk to facilitate the delivery of secretory vesicles to the plasma membrane during polarized exocytosis [12]. Therefore, we examined whether protein secretion would be affected in Mofim1 grown under axenic culture conditions compared to the WT. The Mofim1 mutant exhibited a >50% reduction in protein secretion compared to the WT (S9 Fig), indicating that exocytosis was severely affected in this mutant. Mass spectrometry (MS) analysis of these secreted proteins identified more than one hundred proteins, most of which were down-regulated in the Mofim1 mutant. Of these, the amount of the effector AVR-Pia secreted by the Mofim1 mutant was 0.43-fold that secreted by the WT, indicating that MoFim1-mediated actin assembly is required for secretion of effectors (S1 Table).
To investigate which step was impaired during polarized exocytosis, we introduced Snc1-GFP (Snc1, a putative vesicle-bound v-SNARE protein) into the WT and Mofim1 mutant backgrounds. In the WT, Snc1-GFP-labeled vesicles were actively transported to the growing hyphal tip (Fig 8A, and S9 Movie). However, in the Mofim1 mutant, the vesicles moved to the subapical region of the hyphae but failed to move forward to the hyphal tip. The vesicles usually formed aggregates at this site ( Fig 8A and S9 Movie). These results indicate that polarized vesicle exocytosis was impaired in Mofim1.
Next, we introduced Myosin5-GFP fusion protein into the WT and Mofim1 mutant backgrounds. Class V myosin is an actin-based motor responsible for the short-distance transport of vesicles to the Spk and their delivery to specific regions of the hyphal apex [36][37][38]. Myo-sin5-GFP appeared as a bright spot at the growing tips of WT hyphae. By contrast, in Mofim1, no signal was detected at the hyphal tip. Instead, some small spots appeared at the cell membrane ( Fig 8B and S10 Movie). These findings indicate that the actin motor protein Myosin5 is distributed in a disorderly fashion in Mofim1. We also used septin3-GFP to label the Spk. In agreement with the results from FM4-64 staining, no septin3 accumulated in the Spk area in Mofim1 (Fig 8C and S11 Movie). Finally, we observed the expression of the exocyst component Exo84, and the membrane-bound t-SNARE Sec9 in the WT and Mofim1. Both proteins localized to puncta at the cell apex in WT. In the mutant, the distribution pattern of Exo84 proteins was similar to that in the WT. The membrane-bound t-SNARE protein Sec9 was detected at the cell membrane, although its distribution pattern was diffuse (Fig 8D and 8E, S12 Movie). These results indicate that without MoFim1, the activity of the septin-dependent exocyst is compromised prior to tethering secretory vesicles to the plasma membrane.

MoFim1 is required for invasive hyphal growth and expansion in rice cells
The Mofim1 mutant exhibited disrupted conidiation, preventing us from investigating its intracellular behavior in plant cells. Therefore, we generated a MoFim1 RNAi line to knock down MoFim1 expression. Three DNA fragments (~200 bp from the 5 0 terminus, the middle

PLOS PATHOGENS
Fimbrin organizes actin assembly in M. oryzae region, and the 3 0 terminus of the MoFim1 open reading frame) were used for RNA hairpin construction (named RNAi-1, RNAi-2, and RNAi-3, respectively). We transformed these RNAi constructs into WT M. oryzae. The RNAi construct from the 5 0 terminus (RNAi-1) exhibited good efficiency for silencing MoFim1. The growth rate of the RNAi-1 line was comparable to that of the Mofim1 knockout mutant, in contrast to RNAi-2 and RNAi-3 ( Fig 9A). Although conidiation was dramatically reduced in RNAi-1 (~20% of the WT rate), the RNAi-1 M. oryzae spores did not exhibit great changes in morphology compared to the WT (Fig 9B). When we examined appressorium formation on hydrophobic glass surfaces, after 8 h of induction, a circular appressorium labeled with Lifeact-GFP appeared in the WT (Fig 9C). However, in RNAi-1, the appressoria were usually deformed (Fig 9D and 9E).
The RNAi-1 spores usually developed long hyphae outside of the rice epidermal sheath cells and only a small portion could penetrate (S10 Fig). In invasive WT hyphae, an actin cytoskeleton was usually present in the hyphal tip or at the narrow site in which the hyphae penetrated into the neighboring cell (left panels of Fig 9F-9H). By contrast, invasive RNAi-1 hyphae rarely exhibited this type of actin organization (right panels of Fig 9F-9H). Most invasive hyphae of RNAi-1 were confined to the first penetrated cell (102 cells were examined). By contrast, 80% of WT invasive hyphae (105 cells were examined) penetrated into the other cells, but onlỹ 12% of those in RNAi-1 (118 cells were examined) expanded to the neighboring cells at 36 h after infection (Fig 9H). These results indicate that MoFim1 is required for both the initial plant cell penetration and the expansion of hyphae from one plant cell to another.

Host-induced gene silencing of MoFim1 improves rice blast resistance
Host-induced gene silencing (HIGS) is a powerful strategy for developing transgenic rice cultivars to control fungal diseases and is a useful tool for investigating gene function in pathogens [39,40]. To test HIGS with MoFim1, we generated transgenic rice lines (O. sativa ssp. japonica cv. Nipponbare, susceptible to M. oryzae strain Y34) overexpressing artificial small interfering RNAs (siRNAs) produced from the same DNA fragment used for construction of RNAi-1. We also generated transgenic lines for HIGS of GFP as a parallel negative control. Southern blot analysis showed that the three selected T 2 HIGS-MoFim1 transgenic lines were all single-copy insertions (S11 Fig). Specifically, WT M. oryzae spores subjected to dual fluorescence labeling with GFP and mCherry were used in the rice leaf sheath penetration assays. After 72 h of infection, GFP/mCherry-labeled M. oryzae successfully penetrated into WT rice sheath cells. Strong GFP and mCherry signals from M. oryzae were observed in the plant cells (Fig 10A). We also incubated GFP/mCherry-labeled M. oryzae spores with HIGS-GFP transgenic rice sheath cells. While fluorescent mCherry signals were not obviously altered in these cells, the intensity of GFP signals was significantly reduced in M. oryzae that had penetrated into plant cells ( Fig  10B and 10E). qRT-PCR analysis confirmed the silencing of GFP expression in HIGS-GFP transgenic rice cells (Fig 10H), verifying that the HIGS system functioned efficiently.
We used Lifeact-GFP/MoFim1-mCherry double-labeled M. oryzae to infect sheath cells from WT and HIGS-MoFim1 transgenic rice plants. In WT sheath cells, Lifeact-GFP-labeled actin accumulated in some invasive hyphal tips or at the narrow sites in which the hyphae penetrated into neighboring cells (as observed in Fig 9). MoFim1-mCherry colocalized with these actin aggregates at these sites ( Fig 10C). However, in HIGS-MoFim1 transgenic rice plants, along with the significantly reduced fluorescence intensity of MoFim1-mCherry, a Lifeact-GFP labeled actin cytoskeleton was not usually detected at the hyphal tip or the narrow hyphal penetration site (Fig 10D and 10F). The reduced expression of MoFim1 was further verified by

PLOS PATHOGENS
Fimbrin organizes actin assembly in M. oryzae qRT-PCR (Fig 10I). In addition, most invasive hyphae were confined to the first penetrated cell, which is reminiscent of the infection defects in RNAi-1 (Fig 9). Finally, when we sprayed WT M. oryzae spore suspensions onto WT and HIGS-MoFim1 transgenic plants, after 5 d of inoculation, the HIGS-MoFim1 plants produced significantly smaller lesions than the controls (Fig 10G and 10J). These findings indicate that HIGS of MoFim1 improves rice resistance against M. oryzae.

M. oryzae forms a unique actin structure that facilitates hyphal growth and infection
The growth of hyphal cells of filamentous fungi occurs at the tip via a process analogous to pollen tube growth and root hair elongation [41]. This process, referred to as tip growth, requires precise regulation of the actin cytoskeleton. Characterizing the various actin structures in these cell types is currently an active area of research. For example, a population of apical actin filaments has been observed at the growing apex of the Arabidopsis pollen tube. These actin filaments, which originate from the apical membrane, form a specialized structure for vesicle transport [34]. Many interconnected, thick actin bundles are present in the shaft of the root hair. However, at the apex of the growing root hair, the actin organization markedly changes, as actin bundles become progressively thinner and a dense array of filamentous actin is present [42]. Yu et al. (2019) recently investigated the organizing principles and dynamic behavior of the cytoskeleton in cotton (Gossypium hirsutum) fibers, which elongate via a tip-biased but diffuse growth mode. The elongating cotton fibers assemble a cortical filamentous actin network that extends along the cell axis, ultimately forming actin strands with closed loops in the tapered fiber tip [43]. These findings provide important insights into the cellular mechanisms of tip growth during cell elongation.
In the filamentous fungi, dynamic assembly of the actin cytoskeleton forms diverse actin structures in development and pathogenesis. Considerable effort, using actin-binding protein or peptide (Lifeact) labeling, has shown that the actin cytoskeleton assembles as patches, cables, and rings in filamentous fungi. In Neurospora crassa, the actin cytoskeleton forms a subapical collar of actin patches and a concentration of actin within the core of the Spk [44][45][46][47]. In actively growing Aspergillus nidulans hyphae, the actin assembles as patches, which are highly mobile throughout the hyphae and are concentrated near hyphal apices [30]. A subapical actin web was observed at several microns behind the growth site [48]. In M. oryzae, the actin cytoskeleton is thought to play crucial roles in development and pathogenesis [19,20,49]. A toroidal F-actin ring assembles in the appressorium and is required for the protrusion of a rigid penetration peg to breach the leaf surface [21]. These findings lay the foundation for investigating the roles of the actin cytoskeleton in M. oryzae infection. The M. oryzae MAPK Pmk1 was recently found to function upstream of the actin cytoskeleton, regulating its organization during the infection of adjacent plant cells [22]. However, despite the importance of the actin cytoskeleton, our understanding of how the actin cytoskeleton functions is limited, and we lack a unified view of the organization and dynamics of actin filaments, especially during hyphae polar growth and expansion in plant cells. In this study, through super-resolution live-cell

PLOS PATHOGENS
Fimbrin organizes actin assembly in M. oryzae imaging, we found that M. oryzae (like the fungus Neurospora crassa) produced a subapical collar of actin patches and intense actin at the Spk (Fig 2). In addition, we detected a dense assembly of actin filaments that were continuously generated from the Spk and extended (bỹ 5 μm) into the subapical area of the hyphal tip (Fig 2B, S1 Movie). These actin filaments between the Spk and subapical actin collar have not been clearly observed in other fungi such as Neurospora crassa, Aspergillus nidulans, or Colletotrichum graminicola [30,[44][45][46][47][48]50]. We further noticed that these F-actin filaments radiated from the Spk and were connected to the cell membrane, forming bowl-shaped structures capping the hyphal tips (Fig 2B-2D). Since the delivery of endocytosed vesicles to the Spk area is dependent on the actin cytoskeleton (Figs 7 and 8A), it appears that the endocytosed vesicles, as well as the polarisome, are delivered to the subapical region and transported along F-actin cables to the Spk. Thus, based on our observations, we identified a unique organization pattern of actin at the hyphal tip unlike those observed in plant pollen tubes, root hairs, or cotton fibers.
A previous study revealed that F-actin also assembled in invasive hyphae and the actin cables might play an essential role in plant infection [50]. Until now, our knowledge of the dynamic organization of actin in invasive hyphae has remained limited. We determined that the actin cytoskeleton also accumulates in some invasive hyphal tips (but not in bulbous invasive hyphae) (Figs 4B, 9F-9H and 10C), indicating that these cells undergo polarized growth. Although obvious accumulation of actin at the apex of the invasive hyphae was detected ( Fig  9F-9H), fine F-actin filaments were not frequently observed in these hyphae, like those in vegetative hyphal tips (Fig 4A and 4B). This perhaps was due to the stress environment in plant cells, which affects F-actin formation.
In addition to the hyphal tip, the actin cytoskeleton accumulated at the narrow sites in which hyphae penetrate into the neighboring cells (Fig 9H). In line with our results, the actin cytoskeleton scaffold protein Septin5 assembles at the cell wall contact points, highlighting the organization of the actin cytoskeleton at this region [22]. Mutations of the MAPK Pmk1 affect the location of Septin5 at this site, thus impairing cell-to-cell invasion by M. oryzae [22]. Based on these and the current observations, we speculate that when hyphae invade the adjacent plant cells, the actin in the hyphal tip rapidly becomes organized to build polarity to point towards the plasmodesmata and to cause the hyphae to become narrower, thereby helping them cross into neighboring cells.

MoFim1 organizes the actin cytoskeleton for hyphal tip growth and polarized exocytosis in M. oryzae
Fimbrins are well-characterized actin bundling proteins that function through their two conserved ABDs [51]. In vitro analysis of the biochemical properties of MoFim1 revealed that MoFim1 directly binds to and bundles actin filaments (Fig 5C, and S6 Fig), pointing to the conserved activities of this protein in actin assembly. Vertebrate fimbrins also contain an Nterminal calcium-binding domain consisting of two EF-hand-like motifs that precede two ABDs, but fimbrins in yeast and plants lack this motif [25]. The functional significance of these EF-hand-like motifs is unclear, although they might play a role in calcium binding [52]. In this study, in vivo and in vitro experiments showed that the EF motif alone could not bind to or bundle actin filaments (Figs 5E and 6B). Instead, the ABD1 and ABD2 motifs could bind to but not bundle actin filaments (Figs 5F, 5G, 6C and 6E). The EF motif could form homodimers, facilitating the actin binding/bundling activities of ABD1 and ABD2 (Figs 5H, 5I and  6D-6F, S7 Fig). These findings shed light on how MoFim1 binds to and bundles actin filaments through its various domains.

Fimbrin organizes actin assembly in M. oryzae
The Spk in the hyphal apex is composed of a dense group of vesicles, ribosomes, microtubules, actin, and an amorphous or granular material of undefined nature [7,11]. The Spk is thought to be associated with hyphal tip growth and vesicle transport, which was further verified by our drug treatment experiments (Fig 7B). Although the Spk is a highly dynamic structure maintained at the hyphal tip that moves during hypha growth, we still do not know exactly how the Spk forms or how vesicles are transported from the inner cytoplasm to the hyphal tip [15]. Here, we found that MoFim1 localized to both the actin at Spk and actin cables in M. oryzae hyphae (Fig 4A-4E). Knockout of MoFim1 impaired the formation of both the actin at Spk and the actin cables, thus compromising hyphal tip growth (Fig 3). We noted that the Mofim1 mutant could form an Spk and establish the polar growth sometimes, but it cannot maintain the Spk and polar growth (Fig 3B). Therefore, we speculate that the knockout of MoFim1 mainly prevents the maintenance of polar growth of hyphae. It appears that MoFim1 plays important roles in the establishment of actin structures originating from the Spk involved in hyphal tip growth in M. oryzae.
Polarized exocytosis is an essential process in fungi and is required for cell growth and pathogenesis [12]. This process involves the polarized trafficking of secretory vesicles to the Spk and their subsequent delivery to specific domains of the plasma membrane, where they fuse to provide the enzymes and material needed to build the cell wall and deliver proteins for virulence [7]. The role of the actin cytoskeleton in this continuous process has not been clearly elucidated. In the Mofim1 mutant, the actin cytoskeleton was disorganized at the hyphal tip, indicating that vesicle delivery tracks were disordered (Fig 3). Thus, the subsequent transport of the v-SNARE protein Snc1 and the actin motor protein Myosin5 was disrupted (Fig 8), and protein secretion was compromised (S9 Fig and S1 Table). Based on these findings, we conclude that MoFim1 organizes the actin cytoskeleton in the hyphal tip, providing the tracks for polarized exocytosis. These important insights shed light on the molecular mechanism underlying how secretory vesicles are trafficked to the hyphal tips.

MoFim1 represents a target for M. oryzae control
Rice blast caused by M. oryzae is the most devastating fungal disease of rice [53]. Therefore, it is critical to develop effective means to control this disease. Numerous rice resistance (R) and defense-regulator (DR) genes conferring resistance to M. oryzae have been identified. For example,~25 blast R genes, including the recently characterized Pigm gene from rice cultivar Gumei4, confer blast resistance [54,55]. In addition, multiple DR genes that function in blast resistance have been identified, such as bsr-d1, bsr-k1, spl11, spl33, IPA1, and so on [56][57][58][59][60]. These important R and DR genes have potential applications for breeding rice cultivars with enhanced M. oryzae resistance.
In addition to these rice R or DR genes, key genes that function in M. oryzae development or pathogenesis could serve as targets for the control of fungal diseases via HIGS [39,61]. Here, we investigated the physiological roles of MoFim1 and found that this protein plays crucial roles in actin cytoskeleton organization in M. oryzae during development and virulence. Moreover, transgenic rice produced in our HIGS experiments exhibited strong resistance to M. oryzae (Fig 10). This resistance was most likely due to the downregulated expression of MoFim1 (Fig 10I), which would lead to disorganization of the actin cytoskeleton in the invasive hyphae, thereby compromising hyphal tip growth and expansion in rice cells (Fig 10C and  10D). Since this fimbrin gene is highly conserved in M. oryzae, it might represent an ideal target for M. oryzae control.

PLOS PATHOGENS
Fimbrin organizes actin assembly in M. oryzae

Plant materials and Magnaporthe oryzae culture
All M. oryzae isolates used in this study were derived from the WT strain Y34 (kindly provided by Prof. LiHuang Zhu, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences). All strains were cultured on CM agar plates for growth and SRB medium for conidia production. Liquid CM was used to prepare the mycelia for DNA and RNA extraction. Y34 susceptible rice (Oryza sativa ssp. japonica cv. Nipponbare) was used for disease and transgenic analyses.

Targeted MoFim1 deletion and plasmid construction
The deletion of MoFim1 in M. oryzae was generated using the standard one-step gene replacement method [35]. To obtain the Mofim1 complemented strain, the sequence containing the MoFim1 gene and 1.5-kb native promoter region was ligated to mCherry and cloned into the pKNTG binary vector.
To construct plasmids expressing Snc1-GFP, Myosin5-GFP, Sep3-GFP, Exo84-GFP, and Sec9-GFP, the related DNA fragments from the M. oryzae genome and the~1.5-kb native promoter region were amplified and cloned into the pKNTG binary vector. To construct the Lifeact-GFP plasmid, the Lifeact DNA sequence was ligated to GFP and cloned into the PsulPH vector [62,63].
The HIGS experiments mainly followed the method reported previously [39,40]. The genespecific DNA fragments (about 200 bp) for MoFim1 and GFP were cloned into the plant RNAi pANDAHK35 vector in opposite orientations on either side of a GUS linker. Then the fragment containing the gene-specific DNA fragments in two opposite orientations and the GUS linker were amplified and cloned into the pCAMBIA1300 vector and the recombinant plasmids were used for rice transformation. All constructs were generated via homologous recombination cloning (ClonExpress MultiS One Step Cloning Kit, Vazyme Biotech, C112); all primers with restriction enzyme sites are listed in S2 Table. The recombinant plasmids were transformed into M. oryzae protoplasts as described previously [35].

Pathogenicity assay
The M. oryzae WT, Mofim1, and complemented strain were cultured on SRB medium for 7 d. As Mofim1 could not generate spores, the same region of the culture plate for all three strains was used to infect punched rice leaves (the third leaf of each seedling, cv. Nipponbare). The disease phenotypes of the leaves were observed and imaged at 5 days post infection (dpi).
For spray inoculation of conidia, a conidial suspension (1 × 10 5 conidia/ml) was sprayed onto rice leaves (WT and HIGS transgenic plants) with a sprayer. Inoculated plants were grown in a growth chamber at 28˚C with high humidity in the dark for the first 24 h, followed by a 12-h/12-h light (20,000 lux)/dark cycle [64].
To infect rice sheath cells, M. oryzae spores (1 × 10 5 conidia/ml) from WT and the MoFim1 RNAi strains were diluted in a 0.2% (w/v) gelatin solution. The inner leaf sheath cuticle cells of 3-week-old rice plants were inoculated with the conidial suspension and incubated under humid conditions at 28˚C.

Measurement of the hyphae growth rate
The hyphae of the WT, Mofim1 mutant, and the complemented strains were cultured in liquid CM overnight. The hyphae of the WT, Mofim1 and the complemented strains were observed

PLOS PATHOGENS
Fimbrin organizes actin assembly in M. oryzae under a microscope (Zeiss LSM880, with a 20x objective). Time-lapse imaging was conducted and the hypha's growth rate was calculated according to the growth distance over 60 minutes.

Yeast two-hybrid assay
To validate homodimer formation of the EF domain, a yeast two-hybrid assay was performed using the Matchmaker Yeast Two-Hybrid System (Clontech) following the manufacturer's instructions. The cDNA encoding the EF domain was cloned into the AD and BD vectors. The recombinant plasmids were cotransformed into AH109 cells and their growth examined on DDO and QDO medium. The pGADT7-T/pGBKT7-53 (AD-T/BD-53) plasmid was used as a positive control, and pGADT7/pGBKT7 (AD/BD) was used as a negative control.
In vitro protein purification and F-actin binding/bundling assay cDNAs from MoFim1 and its various domains (shown in Fig 5A) were fused to GFP and cloned into the pET28a vector to produce His-tagged MoFim1 and MoFim1-GFP, EF-GFP, ABD1-GFP, EF-ABD1-GFP, ABD2-GFP and EF-ABD2-GFP fusion proteins. All primers and restriction enzyme sites are listed in S2 Table. All constructs were transformed into Escherichia coli strain BL21 (DE3), and recombinant protein expression was induced by the addition of 0.5 mM isopropyl-β-D-thiogalactoside at 16˚C overnight. Recombinant proteins were purified following the manufacturer's instructions.
The high-and low-speed co-sedimentation assays were conducted as previously described [25]. F-actin was prepared from rabbit muscle G-actin proteins as previously reported [26]. Briefly, G-actin (4 μM) was incubated at 22˚C for 60 min alone or with 0 to 8 μM MoFim1 in KMEI buffer (10× stock: 500 mM KCl, 10 mM MgCl 2 , 10 mM EGTA, and 100 mM imidazole, pH 7.0). The samples were centrifuged at 200,000 g for 60 min for the high-speed co-sedimentation assay or 13,500 g for 30 min for the low-speed co-sedimentation assay. The proteins in the supernatants and pellets were separated by SDS-PAGE and visualized by staining with Coomassie Brilliant Blue R 250.
Visualization of actin filaments in the presence of MoFim1 and its domains was performed by fluorescence microscopy as reported previously [65]. Prepolymerized F-actin (1 μM) was incubated with MoFim1 and various truncated proteins (1 μM) at room temperature for 30 min and labeled with Alexa561-phalloidin (Thermo Fisher, A12380). The images were obtained under a confocal microscope (Zeiss LSM880 Airyscan) at 488 or 561 nm.

Extraction of secreted proteins from M. oryzae mycelia
Secreted proteins were extracted from M. oryzae mycelia as described previously [12]. Fresh WT and Mofim1 mycelia were cultured in liquid CM for 48 h. An equal weight of WT and Mofim1 mycelia was harvested by filtration and transferred to liquid GMM for 24 h. The secreted proteins in the medium were collected and condensed in an ultrafiltration tube (3 kD, Millipore). The protein samples were quantified by the Bradford method.

PLOS PATHOGENS
Fimbrin organizes actin assembly in M. oryzae

RNA extraction and qRT-PCR analysis
Total RNA was extracted from the samples using a Total RNA Purification kit (TransGen, ET101-01) according to the manufacturer's protocol. qRT-PCR analysis of GFP and MoFim1 in transgenic rice (generated by HIGS) was performed using SYBR Green Real-time PCR Master Mix (Toyobo, Japan), with M. oryzae Histone used as the internal control. All reactions were conducted in triplicate using the primers shown in S2 Table.

Observation of fluorescent signals by super-resolution live-cell imaging
Live-cell imaging was conducted under a super-resolution confocal microscope (Zeiss LSM880) equipped with an Airyscan detector. The Airyscan unit acquires data simultaneously using 32 detectors arranged in a hexagonal array [67]. The images were processed and analyzed using ImageJ (http://rsbweb.nih.gov/ij), as described previously [68]. F-actin skewness and density were measured as previously described [69]. The maximum projection of the image stack was used to record the global organization of the actin cytoskeleton in fast-growing M. oryzae hyphal cells.

Southern blot analysis
For MoFim1 deletion verification, Kpn I was used to digest the genomic DNA from the WT and the Mofim1 mutant. The digested products were separated, blotted to a membrane, and hybridized with the two indicated biotin-labeled probes (S3 Fig). The probe was designed according to the disruption strategy and was amplified from genomic DNA. To confirm HIGS-MoFim1 transgenic rice, the biotin-labeled probe was used to hybridize to BamH Idigested genomic DNA from the WT and the transgenic rice plants. All the primers used in the Southern blot experiments are listed in S2 Table. The detection was carried out according to the manufacturer's instructions (Thermo DNA Detection kit, 20148).

Statistical analysis
Skewness analysis was performed to quantify the extent of actin bundling in hyphae according to a previously described method [70,71]. The z-series stacks of all optical sections were filtered using Gaussian blur to reduce background noise and then skeletonization was assessed with ThinLine, a JAVA plug-in procedure [70]. The actin filament pixels were collected into a single image using maximum intensity projections and the skewness values were calculated. The actin filament counts in hyphae, the area of the diseased rice leaves, and the fluorescence intensity of the images were obtained using the ImageJ/Fiji platform (http://rsbweb.nih.gov/ ij). Then two-tailed t-tests were used to determine the significance of results. The numerical data and statistical analysis that were used to generate graphs were provided in S1 Data.

PLOS PATHOGENS
Fimbrin organizes actin assembly in M. oryzae