Volume 222, Issue 4 p. 1909-1923
Full paper
Free Access

Plant pathogenic fungi Colletotrichum and Magnaporthe share a common G1 phase monitoring strategy for proper appressorium development

Fumi Fukada

Fumi Fukada

Laboratory of Plant Pathology, Life and Environmental Sciences, Graduate School of Kyoto Prefectural University, Sakyo, Kyoto, 606-8522 Japan

Search for more papers by this author
Sayo Kodama

Sayo Kodama

Laboratory of Plant Pathology, Life and Environmental Sciences, Graduate School of Kyoto Prefectural University, Sakyo, Kyoto, 606-8522 Japan

Search for more papers by this author
Takumi Nishiuchi

Takumi Nishiuchi

Division of Functional Genomics, Advanced Science Research Centre, Kanazawa University, Kanazawa, 920-0934 Japan

Search for more papers by this author
Naoki Kajikawa

Naoki Kajikawa

Laboratory of Plant Pathology, Life and Environmental Sciences, Graduate School of Kyoto Prefectural University, Sakyo, Kyoto, 606-8522 Japan

Search for more papers by this author
Yasuyuki Kubo

Corresponding Author

Yasuyuki Kubo

Laboratory of Plant Pathology, Life and Environmental Sciences, Graduate School of Kyoto Prefectural University, Sakyo, Kyoto, 606-8522 Japan

Author for correspondence:

Yasuyuki Kubo

Tel: +81 075 703 5613

Email: [email protected]

Search for more papers by this author
First published: 04 February 2019
Citations: 22

Summary

  • To breach the plant cuticle, many plant pathogenic fungi differentiate specialized infection structures (appressoria). In Colletotrichum orbiculare (cucumber anthracnose fungus), this differentiation requires unique proper G1/S phase progression, regulated by two-component GTPase activating protein CoBub2/CoBfa1 and GTPase CoTem1. Since their homologues regulate mitotic exit, cytokinesis, or septum formation from yeasts to mammals, we asked whether the BUB2 function in G1/S progression is specific to plant pathogenic fungi.
  • Colletotrichum higginsianum and Magnaporthe oryzae were genetically analyzed to investigate conservation of BUB2 roles in cell cycle regulation, septum formation, and virulence. Expression profile of cobub2Δ was analyzed using a custom microarray.
  • In bub2 mutants of both fungi, S phase initiation was earlier, and septum formation coordinated with a septation initiation network protein and contractile actin ring was impaired. Earlier G1/S transition in cobub2Δ results in especially high expression of DNA replication genes and differing regulation of virulence-associated genes that encode proteins such as carbohydrate-active enzymes and small secreted proteins. The virulence of chbub2Δ and mobub2Δ was significantly reduced.
  • Our evidence shows that BUB2 regulation of G1/S transition and septum formation supports its specific requirement for appressorium development in plant pathogenic fungi.

Introduction

To initiate plant infection, many plant-pathogenic fungi develop an infection structure called the appressorium to penetrate the plant epidermis (Bourett & Howard, 1990). Colletotrichum sp. and Magnaporthe oryzae are two such pathogens that differentiate a dome-shaped or lobed appressorium that is separated from the germ tube by a septum (de Jong et al., 1997; Perfect et al., 1999). In a multistage hemibiotrophic infection strategy, the appressoria breach the host cuticle and cell wall using enormous turgor pressure and enzymatic degradation, then differentiate a narrow penetration peg (Talbot, 2003). Inside the host, infection hyphae invaginate the plasma membrane without penetrating it. In this initial biotrophic phase of fungal growth, the plant cells are still alive. The fungus then switches to a necrotrophic phase and differentiates thin, fast-growing hyphae that kill and destroy host tissues (Talbot, 2003). Because these species are included among the top 10 scientifically and economically important fungal pathogens (Dean et al., 2012), understanding the mechanisms involved in the infection strategy of these species is critical for targeting the earliest stages of plant infection and better controlling disease.

Such infection-related fungal morphogenesis is coordinated through proper progression through the cell cycle and cell division (Flor-Parra et al., 2006; Saunders et al., 2010a). In M. oryzae, early appressorium development requires the DNA in the nucleus of the germinating conidium to replicate (Saunders et al., 2010a; Osés-Ruiz et al., 2017). In the corn smut fungus Ustilago maydis, appressorium development is required for G2 arrest in the cell cycle in the dikaryotic infective filament cells (Castanheira & Pérez-Martín, 2015). For Colletotrichum orbiculare, our previous work demonstrated that appressorium formation is required for proper cell cycle progression from G1 to S, which is regulated by a two-component GTPase activating protein (GAP) complex (CoBub2 and CoBfa1), which targets GTPase CoTem1 (Fukada & Kubo, 2015). Disruption mutants in either GAP component initiate earlier G1/S progression during conidial germination and, consequently, earlier nuclear division and atypical binucleation. This earlier G1/S transition in GAP mutants results in reduced virulence since these mutants do not form a penetration peg. This failure of penetration depends on an attenuated assembly of septin and actin at the appressorium pore and enhanced plant defense responses such as callose deposition and reactive oxygen species accumulation at sites the appressorium attempts to penetrate. Thus, the lack of GAP component CoBub2/CoBfa1 results in earlier G1/S progression during appressorium formation, and the appressorium is not functional because its infection peg cannot penetrate the host. However, it remains unclear what gene sets are regulated by Bub2 to control G1/S progression and why cobub2Δ impaired the penetration step and enhanced plant defense response.

Intriguingly, the amino acid sequences of the GAP complex CoBub2/CoBfa1 and its GTPase CoTem1 are well conserved from yeast to humans; however, this signal cascade has diverse functions depending on the species (Bardin & Amon, 2001; McCollum & Gould, 2001; Johnson et al., 2012). For example, Bub2 and Bfa1 in Saccharomyces cerevisiae negatively regulate the mitotic exit network (MEN) pathway, which controls mitotic exit and cytokinesis (Krishnan et al., 2000). In contrast, Cdc16 and Byr4, homologues of Bub2 and Cdc16 in Schizosaccharomyces pombe, negatively regulate the septation initiation network (SIN) pathway, which triggers septum formation rather than mitotic exit (Furge et al., 1998). The MEN/SIN pathway is initiated by activation of the GTPase Tem1/Spg1, and, subsequently, Tem1/Spg1 activates downstream protein kinases and their associated subunits (Balasubramanian et al., 1998; Jin et al., 2006; Hachet & Simanis, 2008). In S. pombe, the SIN pathway comprises three protein kinase modules: Cdc7, Sid1-Cdc14, and Sid2-Mob1 (Fankhauser & Simanis, 1993; Sparks et al., 1999; Guertin et al., 2000; Hou et al., 2000). Loss-of-function mutations in SIN activators (e.g. spg1, cdc7) lead to the formation of elongated cells with multiple nuclei (Nasmyth & Nurse, 1981; Schmidt et al., 1997; Balasubramanian et al., 1998), whereas SIN inhibitor mutants (cdc16 or byr4) develop multiseptated cells that fail in cell cleavage (Minet et al., 1979; Song et al., 1996). In S. pombe and the filamentous fungi Neurospora crassa and Aspergillus nidulans, the SIN pathway is critical for selecting the division site, spatiotemporal organization of the contractile actomyosin ring (CAR), and coordination of cell cycle progression with CAR constriction (Roberts-Galbraith & Gould, 2008). In these species, the homologues of BUB2 negatively regulate SIN through inactivating GTPase Tem1/Spg1 so that cytokinesis and septum formation are not induced during interphase (Simanis, 2003; Heilig et al., 2013). A nuclear Dbf2-related (NDR) kinase Sid2p encodes an NDR kinase, and its associated subunit Mob1p localizes at the spindle pole body. At the end of anaphase, as the spindle breaks down, Sid2p-Mob1p is relocalized at the site of the CAR and remains there during ring contraction, flanking the developing septum, thus functioning as an important player in SIN (Sparks et al., 1999; Hou et al., 2000; Salimova et al., 2000). So far, the MEN/SIN pathway and its inhibitor Bub2-Bfa1/Cdc16-Byr4 have been characterized in yeasts, nonpathogenic filamentous fungi, and mammals; however, their function in G1/S progression has been reported only for C. orbiculare, and cobub2Δ or cobfa1Δ were not defective in septation, implying Bub2 and Bfa1 are not important for septation in C. orbiculare.

In hemibiotrophic plant pathogens, a study of M. oryzae showed a temperature-sensitive mutant of SEP1, a homologue of S. pombe Cdc7, a downstream target of Tem1/Spg1, displayed enhanced septation and multinucleation at the restrictive temperature and impaired the development of a functional appressorium (Saunders et al., 2010b). This observation suggests the component of the MEN/SIN pathway could be used for nuclear division as well as septum formation, and regulation of these is required for successful plant infection in M. oryzae. This observation led us to wonder whether the function of Bub2 is conserved in plant pathogenic fungi in respect of G1/S progression as well as septum formation and the connection between regulation of these and development of a functional appressorium.

Although both Colletotrichum sp. and M. oryzae are hemibiotrophic, the septum formation during conidiation, germination, and appressorium development is different. In most of the Colletotrichum sp., such as Colletotrichum higginsianum, Colletotrichum graminicola, and Colletotrichum gloeosporioides, aseptate conidia become septate following mitosis, just before germination (Damm et al., 2014). By contrast, C. orbiculare and related species remain aseptate during germination (O'Connell et al., 1992; Damm et al., 2013). M. oryzae first produces single-celled conidiogenous cells from the conidiophore. After one round of mitosis and formation of a septum, the tip compartment will then undergo one more round of mitosis and cytokinesis to form three-celled pyriform conidia (Howard & Valent, 1996; Liu et al., 2010). After germination, all these fungi develop an appressorium and the septum is formed at the neck of the developing appressorium. Thus, the regulation of septum formation as well as cell cycle progression could be different depending on the fungal species.

In this study, we investigated the function of Bub2 in G1/S progression and septum formation, adopting two other plant pathogenic fungi, C. higginsianum and M. oryzae, as representative fungal species that differ in their progression through the cell cycle and in the process of septum formation, and analyzed the importance of Bub2 for completing appressorium differentiation and infecting the plant. Furthermore, we analyzed the transcriptome of cobub2Δ to determine which genes contribute to earlier G1/S progression and reduced virulence in cobub2Δ.

Materials and Methods

Fungal and bacterial strains, culture conditions, and transformation

Strain 104-T (MAFF240422) of C. orbiculare, strain 337-5 of C. higginsianum, and strain Hoku-1 of M. oryzae (MAFF 02-35004) were used as the wild-type strains. All fungal strains used in this study are listed in Supporting Information Table S1. C. orbiculare strains were maintained at 24°C in darkness on 3.9% (w/v) BD-Difco potato dextrose agar (PDA; Nippon BD, Tokyo, Japan). C. higginsianum and M. oryzae were maintained at 24°C under black light on oatmeal agar (Nippon BD). Escherichia coli DH5α-competent cells were used as a host for gene manipulation and maintained on Luria–Bertani agar at 37°C. Agrobacterium tumefaciens strain C58C1 was used as the transfer-DNA donor for C. higginsianum and M. oryzae transformation and maintained on Luria–Bertani agar at 28°C, based on a previous protocol (Fukada & Kubo, 2015). For C. higginsianum transformation, hygromycin-resistant transformants were selected on PDA containing 100 μg ml−1 hygromycin B (Fujifilm Wako Pure Chemical Corp., Osaka, Japan). Sulfonylurea-resistant transformants were selected on Czapeck agar (30 g sucrose, 2 g sodium nitrate, 1 g dipotassium phosphate, 0.5 g magnesium sulfate, 0.6 g potassium chloride, 15 g agar, 1000 ml water) containing 12 μg ml−1 chlorimuron ethyl (Maruwa Biochemical, Tokyo, Japan). Geneticin-resistant transformants were selected on PDA containing 50 μg ml−1 G418 (Nakalai Tesque, Kyoto, Japan). For M. oryzae transformation, hygromycin-resistant transformants were selected on PDA containing 100 μg ml−1 hygromycin. Sulfonylurea-resistant transformants were selected on fCzapek agar containing 40 μg ml−1 chlorimuron ethyl. Geneticin-resistant transformants were selected on complete agar medium (5% sucrose, 3% casamino acids, 3% yeast extract, 1.5% agar) containing 2 mg ml−1 G418. All selection media contained 25 μg ml−1 meropenem hydrate (Sumitomo Dainippon Pharma, Osaka, Japan). For ectopic transformation with a fusion of Lac operator repeats, protoplasts were transformed using polyethylene glycol as described previously (Kubo & Furusawa, 1991).

Microscopy

For examining conidial germination and appressorium formation of C. higginsianum and M. oryzae, 10 μl of conidial suspension (105 conidia ml−1) was placed on a cover glass (Matsunami Glass, Osaka, Japan) and incubated in a humid box at 24°C in the dark. For the staining of cellular membrane, diluted FM 4-64 (Thermo Fisher Scientific, Waltham, MA, USA) in distilled water was added to the incubated fungal cell suspension at a final concentration 40 ng μl−1 and incubated 30 min in the dark. Cells were washed twice before imaging. For staining nuclei of living cells, 2 μl of 100 mg ml−1 Hoechst 33342 (Dojindo Laboratories, Kumamoto, Japan) was added to cells on samples on the glass slide and incubated for 10 min. For assessing penetration hyphae formation of C. higginsianum, 5 μl of a conidial suspension (105–106 conidia ml−1) was spotted onto the adaxial surface of leaves of 4-wk-old Aribadopsis thaliana Col-0 plants, which were then incubated in a phytochamber at high humidity. After 3 d, leaves were stained with lactophenol–Trypan blue (10 mg Trypan blue, 10 g phenol, 10 ml lactic acid, 10 ml deionized water; Keogh et al., 1980). Leaves were boiled for 60 s in staining solution, decolorized in chloral hydrate (2.5 g chloral hydrate in 1 ml deionized water) for 48 h, then mounted in deionized water and viewed using bright-field or phase-contrast microscopy. For assessing penetration hyphae formation of M. oryzae at 72 h post inoculation (hpi), leaves were placed in boiling lactophenol (1 : 1 : 1 : 1 lactic acid–phenol–glycerol–distilled water/ethanol, v/v/v/v) for 2 min, then observed with bright-field microscopy.

For observations of nuclear or cellular membrane stained cells, and green fluorescent protein (GFP) or red fluorescent protein (RFP)-labelled strain, fluorescence was detected using a Zeiss Axio Imager M2 Upright microscope with an Axio Cam MRm digital camera (Carl Zeiss AG, Oberkochen, Germany) and excitation/barrier filter set of 470 nm/509 nm for GFP and 595 nm/620 nm for RFP. Images were acquired with a ×100 oil immersion lens (Plan Apochromat) using Axiovision 4.8. The signal from GFP-lacI was observed using ×1.63 intermediate variable magnification. The diameter of appressorial necks was analyzed using ImageJ (National Institutes of Health, Bethesda, MD, USA; http://rsb.info.nih.gov/ij/).

Plant growth and infection

Cucumber (Cucumis sativus L. ‘Suyo’) was grown with a 16 h photoperiod at 24°C for 7 d, and cotyledons were used to assay infection after inoculation with conidia from a 5-d-old culture of C. orbiculare as described by Tsuji et al. (2003). For the virulence assay of C. higginsianum, 10 μl of conidial suspension (1 × 106 conidia ml−1 in 0.02% Tween 20 solution) was spotted onto the adaxial surface of leaves of 6- or 7-wk-old plants of A. thaliana Col-0, then maintained under high humidity in the dark for 24 h, then in a phytochamber at 24°C with a 16 h photoperiod for 6 d. For the virulence assay of M. oryzae, 8-d-old barley (Hordeum vulgare L. ‘Nakaizumi-zairai’) seedlings were each spot-inoculated with 10 μl of conidial suspension (1 × 105 conidia ml−1 in 0.2% gelatin solution). The inoculated seedlings were maintained in a humid box in the dark for 24 h, then grown in a phytochamber with a 16 h photoperiod at 24°C for 4 d. The lesion size on barley leaves with M. oryzae infection was measured using ImageJ.

Plasmid construction

Plasmids were derived from the following binary vectors: pBIG4MRHrev, carrying the hygromycin resistance gene cassette; pBIG4MRSrev, carrying the sulfonylurea resistance gene cassette; and pBIG4MRNrev, carrying the neomycin resistance gene cassette. For plasmid constructions, the In-Fusion HD Cloning Kit (Takara Bio USA Inc., Mountain View, CA, USA) was used. Plasmids used in this study are listed in Table S2.

For generating gene disruption mutants, plasmids in which cloned genes were replaced with the hygromycin resistance gene cassette were constructed. For generating the ChBUB2 disruption vector, a 1.0 kb fragment of the 5′ upstream region and a 1.0 kb fragment of the 3′ downstream region were amplified from C. higginsianum genomic DNA, and a 1.4 kb fragment of the hygromycin resistance gene was amplified from pBIG4MRHrev. These three fragments were inserted into a linearized pBIG4MRSrev. The same procedures were used to generate the MoBUB2 disruption vector.

For complementation assays of the mutants, plasmids that contained the complete gene, native promoter, and terminator were constructed. To generate the ChBUB2 complement vector, a ChBUB2 fragment that contained 1.0 kb of the 5′ upstream region, the open reading frame, and 1.0 kb of 3′ downstream region was amplified from C. higginsianum genomic DNA. The fragment was inserted into a linearized pBIG4MRSrev. The same procedures were used to construct the MoBUB2 complementation vector and inserted into linearized pBIG4MRNrev.

For chromosome tagging using lac operator/lac repressor system, pCT31S harboring lac operator repeats (Fukada & Kubo, 2015) and pBITEFGFPLacIN containing the GFP-lac repressor fusions were used. For constructing pBITEFGFPLacIN, the TEFp-GFP-LacI-NLS-SV40 poly(A) fragment was amplified from pBITEFGFPLacIB (Fukada & Kubo, 2015), and the fragment was inserted into pBIG4MRNrev.

For visualizing C. higginsianum Dbf2, a plasmid that carries the gene encoding GFP regulated by the native promoter of ChDBF2 was constructed. For constructing the ChDBF2-GFP fusion gene vector, the ChDBF2 complementation vector that carries the ChDBF2 open reading frame with 1.0 kb of the 5′ upstream region and 1.0 kb of the 3′ downstream region was constructed. The GFP fragment was amplified from pBIglyGFP and inserted at the C-terminal end of ChDBF2 in the complementation vector.

Microarray analysis

For sampling germinating conidia and appressoria in planta, 10 μl of a conidial suspension (5 × 105 conidia ml−1) was spotted onto the abaxial surface of cucumber cotyledons, which were then incubated in a humid box at 24°C. After 1, 2 and 4 h, the lower epidermis of the cotyledons was peeled off. For sampling conidia at 0 h, conidia were harvested with sterile distilled water from twenty 5-d-old cultures on PDA, and the suspension of conidia was centrifuged to make a pellet. All samples were ground in liquid nitrogen, and total RNA was prepared using the Agilent Plant RNA Isolation MiniKit (Agilent Technologies, Santa Clara, CA, USA). Microarray analyses were performed as described previously (Kodama et al., 2017) using the C. orbiculare (8 × 60 000, 13 352 independent probes, Design ID: 060762) oligo microarray, according to the Agilent 60-mer Oligo Microarray Processing Protocol (Agilent Technologies).

Results

BUB2 function in G1/S progression is conserved and is involved in septum formation in C. higginsianum and M. oryzae

To investigate whether the function of BUB2 on G1/S progression is conserved and affects septation in plant pathogenic fungi other than C. orbiculare, we analyzed the BUB2 disruption mutants in C. higginsianum and M. oryzae, which differ in their cell cycle progression and septation process during appressorium differentiation. We identified the ChBUB2 gene (XP_018164413) in C. higginsianum, which putatively encodes a protein of 471 amino acids with 89% sequence identity to C. orbiculare CoBUB2 (ENH83696) (Fig. S1). We generated ChBUB2-disrupted mutants and introduced CoHIS H1:GFP into chbub2Δ to observe nuclear behavior. For visualizing septum formation, we used FM 4-64 to stain the cellular membrane. In the wild-type (WT), aseptate conidia with one nucleus started nuclear division before the conidia germinated after 2 h of incubation on a glass slide (Fig. 1a–c). As the nucleus underwent division, a septum formed at the medial region of the conidia. After the appressorium formed, a second round of nuclear division had started by 5 h of incubation, and a septum formed at the neck of the developing appressoria. In contrast, chbub2Δ started nuclear division earlier; unexpectedly, FM 4-64 staining was eccentric or asymmetrical in conidia, indicating that c. 50% of conidia did not form a complete septum, in striking contrast to disruption of the BUB2 homologue in S. pombe, which induces septum formation (Fig. 1a–c). After appressorium initiation by chbub2Δ, a second round of nuclear division resulted in two nuclei in the conidia and appressoria. Of these appressorium-developing conidia, c. 80% of conidia developed septa at the neck of the developing appressoria, whose outlines are clearly visible, but the rest of the conidia did not.

Details are in the caption following the image
Colletotrichum higginsianum BUB2 regulates G1/S progression and negatively affects septum formation during appressorium development. (a) Time-course series of micrographs of nuclei in conidia and infection structures of C. higginsianum wild-type 337-5 and chbub2Δ during appressorium development. The HISTONE H1-GFP gene fusion vectors were introduced into wild-type and chbub2Δ to visualize nuclei. FM 4–64 was used to stain the plasma membrane of C. higginsianum. White arrows, normal septum in 337-5; orange arrows, abnormal septum in chbub2Δ. DIC, differential interference contrast. (b) Mean percentage (± SE;= 3 biological replicates) of cells with various patterns of nuclear distribution and septum formation. Scoring: one nucleus retained in the conidium (black), one nucleus in the conidium and the other in the appressorium through first round of mitosis (red), one nucleus in the appressorium through second round of mitosis (blue), two nuclei in the appressorium through three rounds of mitosis (yellow). At least 200 conidia were scored in each replication. (c) Mean percentage (± SE;= 3 biological replicates) of cells with two types of septum formation. Scoring: one septum in the middle of conidium (light blue), septum at the neck between conidium and appressorium (light green). At least 200 conidia were scored at each time point. (d) Mean percentage (± SE;= 3 biological replicates) of cells in phase S-G2 in wild-type 337-5 and chbub2Δ cells with two patterns of GFP-tagged Lac Repressor fusion (GFP-LacI) spots. At least 50 conidia were scored at each time. (e) Representative images of Lac operator array (LacO)/LacI-GFP-transformed strains showing one fluorescent spot in one nucleus in putative G1 and closely spaced double spots in one nucleus in putative post S phase or G2 phase. Insets show × 1.65 magnification on the area corresponding to the chromosome where the LacO array is integrated. Bars, (a) 10 μm; (e) 5 μm.

To investigate whether the earlier nuclear division in the chbub2Δ phenotype resulted from earlier G1/S progression, we evaluated the time of entry into the S phase during germination. We used a chromosome tagging system with a fluorescent protein by recruiting a GFP-tagged Lac Repressor fusion (GFP-LacI) to an integrated Lac operator array (LacO) as described in our previous study (Fukada & Kubo, 2015). Briefly, with this system, the time of DNA replication is determined based on large-scale chromatin organization; that is, cells containing pre-replicated DNA (G1 phase) exhibit a single fluorescent spot, whereas cells that have undergone DNA replication (post S phase) harbor two closely spaced fluorescent spots, with one on each separated sister chromatid (Straight et al., 1996). In WT containing LacO/LacI-GFP, a single GFP spot was observed in the nuclei of the conidia at the start of incubation. After 60 min, two GFP spots started to appear in the one-celled conidia (Fig. 1d,e). On the other hand, in chbub2Δ, a single GFP spot was observed at 0 h, and double GFP spots in the nucleus appeared by 30 min of incubation (Fig. 1d,e). These results indicated that the S phase starts after c. 60 min of incubation in WT, but chbub2Δ starts S phase c. 30 min earlier than in WT.

To investigate whether BUB2 function is consistent beyond Colletotrichum sp., we identified the MoBUB2 gene (XP_003710862) in M. oryzae, which has conidia with three cells, each separated by a septum. We generated MoBUB2-disrupted mutants and introduced CoHIS H1:GFP into mobub2Δ to observe nuclear behavior. To visualize the septum formation, FM 4-64 was used to stain the cellular membrane. In WT, the nucleus divided in the apical cell when the appressorium was initiated, in conjunction with septum formation at the neck of the developing appressorium at 4 h of incubation (Fig. 2a–c). In the basal cell and the middle cell, nuclear division and septum formation did not occur during the incubation, and each nucleus was degraded by autophagy after appressorium development. On the other hand, in mobub2Δ, in the apical cell after appressorium initiation, nuclear division started earlier than in WT, but nuclear division also occurred in the basal cell and the middle cell, resulting in binucleate cells (Fig. 2a,b). Different from chbub2Δ, some cells of mobub2Δ became multiseptate; that is, they had one more extra septum in the conidia (Fig. 2a,c). At 24 h of incubation, when the septum formation was visualized by calcofluor white staining, most of the WT cells formed a septum at the base of the appressorium; however, c. 30% of mobub2Δ cells did not (Fig. S2). We then introduced chromosome tagging into WT and mobub2Δ to investigate G1/S progression in M. oryzae. To analyze cellular events in one cell, we scored the basal cell, the middle cell, and the apical cell of the conidium and in the appressorium separately to follow cell cycle progression. In the WT containing LacO/LacI-GFP, only single spots were observed at the nucleus in the nongerminating cells, whereas two GFP spots started to appear in the apical cells after germination, suggesting the S phase starts only in the apical cells (Figs 2d, S3). On the other hand, the two GFP spots appeared earlier in the apical cells, and also in the middle cells and the basal cells of mobub2Δ. After appressorium development, two GFP spots appeared earlier in mobub2Δ, and c. 50% of the appressoria had four GFP spots at 24 h, indicating the nucleus in mobub2Δ entered a second round of the S phase (Fig. S3g). Overall, these results demonstrated that the G1 arrest was released immediately after the cells enter G1 phase in both of chbub2Δ and mobub2Δ and conidia became aseptate in chbub2Δ but multiseptate in mobub2Δ, indicating an inconsistent phenotype with that of the mutants of the yeast homologues.

Details are in the caption following the image
MoBUB2 is required for G1/S progression and contributes to proper septum formation during appressorium development. (a) Representative images of nuclei in Magnaporthe oryzae strain wild-type (WT) Hoku-1, and mobub2Δ with HISTONE H1-GFP introduced after 0, 4, and 8 h incubation. The plasma membrane was stained by FM 4-64. Orange arrows, abnormal septum developed in mobub2Δ. (b, c) Mean percentage (± SE;= 3 biological replicates) of cells with (b) nuclei or (c) septum in WT and the mobub2Δ. At least 100 conidia were scored at each time. (d) Representative images of LacO/LacI-GFP-transformed strains during conidial germination after 3 h incubation of WT and mobub2Δ. Whereas WT has double signals only in the apical cells, mobub2Δ has double signals in each of the three cells. Insets show × 4 magnification on the boxed area corresponding to the chromosome where the LacO array is integrated in each cell in mobub2Δ. Bars, 10 μm.

Localization of SIN component ChDbf2 and filamentous actin assembly to the septation site requires ChBUB2 during appressorium development

Since the frequency of septum formation during conidial germination was reduced in chbub2Δ, we hypothesized that BUB2 in C. higginsianum functions in a previously unknown mechanism for SIN regulation during conidial germination, unlike that in yeast and nonpathogenic filamentous fungi. Therefore, we investigated the relationship among Bub2 and SIN and CAR in C. higginsianum. We identified an S. pombe SID2 homologue in C. higginsianum and named it ChDBF2 (XP_018157267) (Fig. S4). We fused GFP to the C-terminus of ChDbf2 and introduced the construct into WT and chbub2Δ. To monitor CAR formation during septation, we introduced a Lifeact-RFP construct as a filamentous actin (F-actin) marker into WT and chbub2Δ. In WT during conidial germination, ChDbf2-GFP was not detected in interphase, but appeared at the nucleus during the first round of mitosis, predictably associating with the spindle pole body (Fig. S5). After anaphase, ChDbf2-GFP accumulated first in a cortical ring at the cell cortex with uniform signal intensity in the central position between mother and daughter nuclei before the initiation of septum constriction and remained there until septation was complete (Fig. 3a,c). When we observed Lifeact-RFP during conidial germination, it appeared after late mitosis and localized at the medial region of the conidia, and the F-actin ring started to contract in WT (Fig. 3d,f). In contrast, ChDbf2-GFP in chbub2Δ appeared at the nucleus after the first round of mitosis, as in WT, but it localized either eccentrically or asymmetrically in the central position between mother and daughter nuclei (Fig. 3a,c). Lifeact-RFP was also irregularly distributed in chbub2Δ (Fig. 3d,f). After appressorium initiation at 4 h incubation, these signals were found at the neck of developing appressoria as a cortical ring coupled with mitosis in WT; these signals were gone at 6 h (Fig. 3b,c,e,f). On the other hand, in chbub2Δ, the ChDbf2-GFP and Lifeact-RFP signals were observed regardless of the mitotic state and present at the neck of developing appressoria and remained even after completion of appressorium development at 6 h (Fig. 3b,c,e,f). To analyze the ability of the neck of developing appressoria to constrict, we measured the diameter of the neck of appressoria at 24 h incubation. Compared with the neck diameter of WT (1.80 ± 0.03 μm, = 53), the neck diameter of chbub2Δ was somewhat larger (2.22 ± 0.42 μm, = 80), suggesting that the F-actin assembly was unable to contract, resulting in a neck with a larger diameter in chbub2Δ. Thus, these results suggested that ChBUB2 is involved in the proper localization of ChDbf2 and the F-actin assembly to the site of mitotic division and future septation site, and the failure of these localizations in chbub2Δ resulted in the inability to form a conidial septum and forming a clear outlined septum with a larger diameter at the neck of the developing appressorium.

Details are in the caption following the image
ChBUB2 is required for proper ChDbf2 and filamentous-actin localization to future septation site in Colletotrichum higginsianum. (a, b) Representative images of wild-type (WT) 337-5 and chbub2Δ expressing ChDbf2-GFP after anaphase of the first round of mitosis in the conidia at 2 h post inoculation (hpi) (a) and the second round of mitosis in the appressorium developing conidia at 4 hpi (b). Bars, 10 μm. (c) Mean percentage (+SE;= 3 biological replicates) of cells expressing the ChDBF2-GFP gene fusion with three patterns of ChDbf2 localization. Nuclear staining was performed using Hoechst 33342. Scoring: single signals at both nuclei and medial region in the conidium (light gray), single signals at both nuclei but partially distributed in the conidium (dark gray), single signals at both nuclei and the developing appressorium after nuclear migration in appressorium developing conidia (black). At least 100 conidia were scored at each time point. (d, e) Representative images of WT and chbub2Δ expressing Lifeact-RFP after anaphase of the first round of mitosis in the conidia at 2 hpi (d) and the second round of mitosis in the appressorium developing conidia at 4 hpi (e). Nuclear staining was performed using Hoechst 33342. Bars, 10 μm. (f) Mean percentage (+SE;= 3 biological replicates) of cells expressing the LIFEACT-RFP gene fusion with three patterns of Lifeact localization. Scoring: medial region in the conidium (light gray), partially distributed signal in the conidium (dark gray), signal at developing appressorium (black). At least 100 conidia were scored at each time.

BUB2 triggers major changes in expression of genes involved in DNA replication pathway and putative virulence-associated genes in C. orbiculare

Then we wondered what kind of genes are differentially regulated in bub2Δ to start the S phase earlier than WT and reduce virulence on host plants. To identify genes regulated by BUB2, we analyzed the transcriptome of WT and the bub2Δ during appressorium formation using custom microarrays. We used C. orbiculare bub2Δ because the cell cycle and cell morphology are simpler than in C. higginsianum and M. oryzae. We collected samples from four cell cycle stages in WT and cobub2Δ: G1 phase in WT and cobub2Δ (0 h); G1 phase in WT and S or G2 phase in cobub2Δ (1 h); S or G2 phase in WT and M phase in cobub2Δ (2 h); and M phase in WT and second M phase in cobub2Δ (4 h) (Fig. 4a). To monitor the gene expression profile in the fungus on the plant surface, we collected samples at 1, 2 and 4 h after inoculation of C. orbiculare conidia on cucumber cotyledons. For monitoring the gene expression profile before conidial incubation, we harvested fungal conidia from the growing PDA plate (0 h). In cobub2Δ compared with WT, 1676, 1155, 918, and 1059 genes were differentially regulated at 0 h, 1 h, 2 h, and 4 h incubation, respectively (fold-change (FC) ≥ 2; < 0.05). This result shows that c. 7–12% of the 13 479 protein-encoding C. orbiculare genes are differentially regulated by CoBUB2 at each time. From Gene Ontology (GO) enrichment analysis at 2 h incubation, GO terms related to DNA replication such as replication, DNA strand elongation involved in DNA replication were strongly upregulated (< 0.05; Table S3). After 4 h, GO terms related to general cell cycle progression, such as DNA metabolic process, chromosome, and DNA repair, were especially enriched (Table S4), confirming the major differentially regulated event in cobub2Δ is the cell cycle. WT was predicted to possess 564 cell-cycle-associated genes; 194 genes (34%) were differentially regulated in cobub2Δ at the four times (FC ≥ 2, < 0.05). To understand the cell-cycle-related pathway affected by CoBUB2, we annotated the genes differentially expressed between cobub2Δ and WT using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database (http://www.kegg.jp/kegg/pathway.html). A cluster map of the DNA replication pathway, including DNA polymerase complex (e.g. Pola1, Pold1, Pole1), MCM complex (helicase; e.g. Mcm2, Mcm4), replication protein A (e.g. RFA1, RFA2/4), and transcription factor Mbp1 and Swi6, had a higher level of expression at 1 h in WT, whereas the highest expression of these genes in cobub2Δ was at 2 and 4 h (Fig. 4b). The subsequent cell cycle events, such as G2–M transition including a component of the MEN/SIN pathway and anaphase-promoting complex (APC), and the DNA damage checkpoint including Rad17 and Chk1, were upregulated at 4 h in WT. However, in cobub2Δ, MEN/SIN components other than Tem1 were upregulated at 1 and 4 h, suggesting that upregulation of these components leads to mitotic exit. APC-related genes and DNA damage checkpoint genes were constantly upregulated at 1, 2, and 4 h, suggesting APC is important for preventing premature transition from G1 phase to S phase, and the DNA damage response is activated because of the highly activated DNA replication in cobub2Δ. Interestingly, only Tem1 was upregulated at 1 h, whereas other components were upregulated at 4 h in WT, possibly implying only Tem1 acts as an S-phase inducer. These results suggest that, compared with other cell-cycle-related pathways, the DNA replication pathway is the most upregulated immediately after incubation or completion of M phase in cobub2Δ, and the subsequent G2/M and mitotic exit pathways are activated in coordination with cell cycle progression.

Details are in the caption following the image
Select transcript pathways that are activated in cobub2 mutants during appressorium formation in Colletotrichum orbiculare. (a) Schematic representation of the four C. orbiculare cell cycle stages selected for microarray. (b) Expression heat maps of select C. orbiculare genes in wild-type (WT) 104-T and cobub2Δ arranged by their pathway based on the Kyoto Encyclopedia of Genes and Genomes pathway mapper. Shown are pathways for DNA replication, mitotic exit, anaphase-promoting complex (APC), and DNA damage checkpoint. Hierarchical clustering (fold-change (FC) ≥ 2; < 0.05) was used to visualize expression of the C. orbiculare WT genes at 0 h (W0), 1 h (W1), 2 h (W2), and 4 h (W4) and cobub2Δ at 0 h (B0), 1 h (B1), 2 h (B2), and 4 h (B4) after incubation on cucumber leaves. The y-axis shows clustering of C. orbiculare transcripts based on similarity of their expression patterns. Induced differential expression of a gene is indicated by red, and repressed by blue. Abbreviations and transcript identification are as described in Supporting Information Table S4. (c) Number of putative virulence-related genes that were differentially expressed in cobub2Δ compared with WT (FC ≥ 2; < 0.05) in four functional categories. Upregulated genes are indicated in red; downregulated genes are in blue.

To investigate genes that are potentially involved in pathogenesis, we analyzed four gene categories relevant to pathogenicity, including putative small secreted proteins (SSPs, < 300 amino acids long), carbohydrate active enzymes (CAzymes), transcription factors, and membrane transporters in cobub2Δ compared with WT (FC ≥ 2; < 0.05) (O'Connell et al., 2012; Gan et al., 2013). Among 620 genes encoding putative SSPs, 205 genes were differentially expressed in cobub2Δ at 4 hpi (Fig. 4c; Table S5). Of these genes, 121 were downregulated in cobub2Δ, including homologues of C. higginsianum ChEC1 and ChEC5 (Kleemann et al., 2012), Magnaporthe grisea MSP1 (Jeong et al., 2007), and Fusarium oxysporum SIX5 (Lievens et al., 2009). Among 867 genes encoding putative CAzymes, 218 genes were differentially expressed in cobub2Δ at 4 hpi (Fig. 4c; Table S6). Of these genes, 116 were downregulated in cobub2Δ, including 39 enzymes associated with degradation of plant cell wall constituents and four enzymes that bind chitin of the fungal cell wall. These downregulated genes of SSPs and CAzymes in cobub2Δ included 62 genes (51%) or 49 genes (42%) genes, respectively, that were upregulated in planta compared with in vitro at 4 hpi, according to Kodama et al., 2017, suggesting these genes could have important functions for plant infection. Concomitantly, 29 genes encoding plasma membrane transporters, including three of major facilitator super family transporters, and 24 genes encoding transcription factors were downregulated in cobub2Δ (Fig. 4c; Table S7). On the other hand, the upregulated genes in cobub2Δ included homologues of C. higginsianum ChELP2 (Takahara et al., 2016) and C. higginsianum ChEC89 (Kleemann et al., 2012) that are supposed to suppress host defense responses, as well as two genes encoding chitin binding protein (CBM50) and 10 genes encoding glucanase. These upregulated genes might be one explanation for the enhanced callose and reactive oxygen species accumulation with the infection of cobub2Δ. Thus, the deletion of CoBUB2 seems to affect the expression of genes that are important for plant infection, presumably as a secondary effect of cell cycle defect.

Proper cell cycle progression and septum formation regulated by BUB2 are required for pathogenesis by C. higginsianum and M. oryzae

To assess the requirement of BUB2 for fungal pathogenicity in C. higginsianum and M. oryzae, we first observed colony growth on PDA media. Whereas the chbub2Δ showed slight delayed growth compared with WT, mobub2Δ displayed significant growth defect (Fig. S6). To test the ability of the appressorium formation of WT and chbub2Δ, we incubated these strains on glass slides at 24 hpi. The WT typically formed one lobed, melanized appressorium from one conidium and frequently formed an extra-lobed, melanized or nonmelanized appressorium from another side of the conidium. In contrast, chbub2Δ developed only one rounded melanized appressorium or one rounded nonmelanized appressorium (Fig. 5a,b). In a subsequent pathogenicity test of C. higginsianum on leaves of A. thaliana, WT caused severe necrotic lesions; however, chbub2Δ either did not cause lesions or induced only occasional small lesions (Fig. 5c,d). Microscopic observation revealed that > 50% of the appressoria of the WT formed infection hyphae, which invaded the leaf tissue by 3 d after inoculation (Fig. 5e,f). By contrast, chbub2Δ formed very few penetration hyphae. In the case of M. oryzae, the frequency of melanized mature appressoria in mobub2Δ was lower than in WT (Fig. 6a,b); elongated germ tubes and nonmelanized (immature) appressoria formed. In a pathogenicity assay of M. oryzae on barley cv. Nakaizumi-zairai with moderate susceptibility to M. oryzae (Nga et al., 2012), WT caused necrotic lesions, but mobub2Δ did not induce lesions (Fig. 6c). Microscopic observation revealed that > 40% of the appressoria of WT formed infection hyphae, but mobub2Δ formed very few infection hyphae (Fig. 6d). These results indicate that proper cell cycle regulation and septum formation regulated by Bub2 is involved in mature appressorial formation that is necessary for plant penetration in C. higginsianum and M. oryzae.

Details are in the caption following the image
Colletotrichum higginsianum BUB2 is required for infection-related morphogenesis and pathogenesis on Arabidopsis thaliana leaves. (a) Mean percentage (+SE;= 3 biological replicates) of conidia in C. higginsianum wild-type (WT) 337-5, chbub2Δ, and complementation strain chbub2Δ/ChBUB2 (chbub2Δ-C) after 24 h incubation. Scoring: conidia developing two appressoria from each conidium and at least one appressorium is melanized (two melanized), conidia developing only one appressorium that is melanized (one melanized) or is not melanized (one nonmelanized). Appressorium that does not have any pigmentation was counted as a nonmelanized appressorium. At least 150 appressoria were scored for each experiment. (b) Representative images of WT and chbub2Δ with melanized or nonmelanized appressorium after 24 h incubation. (c) Mean percentage (+SE) of lesion size in spot-inoculated A. thaliana Col-0 plants at 7 d after inoculation with C. higginsianum WT, chbub2Δ, and complementation strain. In total, 23 independent plant leaves were measured for each strain from the three biological replicates. (d) Micrographs showing the representative symptom development on Col-0 leaves at 7 d after droplet infection. (e) Mean percentage (+SE;= 3 biological replicates) of appressoria of C. higginsianum WT, chbub2Δ, and complementation strain that formed penetration hyphae. At least 200 appressoria were scored per replication. (f) Micrographs showing penetration hyphae only for WT, not chbub2Δ, on leaf of A. thaliana at 3 d after inoculation. Penetration hyphae were stained with lactophenol Trypan blue. Ap, appressoria; Ph, penetration hyphae. White arrows show nonmelanized appressoria developed from chbub2Δ conidia. Bars, 10 μm.
Details are in the caption following the image
Magnaporthe oryzae BUB2 is necessary for appressorium formation and pathogenesis on barley leaves. (a) Mean percentage (+SE;= 3 replicates) of conidia with melanized appressoria, nonmelanized appressorium, or elongated germ tubes in M. oryzae wild-type (WT) Hoku-1 and mobub2Δ. At least 150 appressoria were scored per replication. ap., appressorium; g. t., germ tube. (b) Representative images of WT with melanized appressorium and mobub2Δ with nonmelanized appressorium or elongated germ tube after 24 h incubation. Bars, 10 μm. (c) Mean percentage (+SE) of lesion size in spot-inoculated barley leaves at 5 d after inoculation with M. oryzae WT, mobub2Δ, or complementation strain (mobub2Δ-C). Conidial suspension was dropped onto detached barley leaves and incubated at 24°C. Total numbers of infected plants from the three biological replicates are indicated above the respective columns. (d) Micrographs showing the representative symptom development on barley leaves at 5 d after droplet infection. (e) Mean percentage (+SE;= 3 biological replicates) of appressoria of M. oryzae WT, chbub2Δ, and complementation strain that formed penetration hyphae. At least 100 appressoria were scored per replication. (f) Micrographs showing penetration hyphae only of WT, not mobub2Δ, on barley leaves 3 d after inoculation. Ap, appressoria; Ph, penetration hyphae. Bars, 10 μm.

To assess the plant immune response in A. thaliana after chbub2Δ infection, first we analyzed callose deposition on A. thaliana WT Col-0 with infection of C. higginsianum WT and chbub2Δ. Compared with the WT infection, no significant differences were observed in chbub2Δ infection (Fig. S7). To investigate the plant immune response by transcriptome analysis, the expression levels of genes involving plant defense response of A. thaliana after inoculation with C. higginsianum WT or chbub2Δ were analyzed at 24, 48, and 72 hpi. The comparative transcriptome analysis with the metabolic pathway of pathogen-associated molecular-pattern-triggered immunity and virulence factors from the KEGG database showed that the expression pattern of these defense genes did not differ significantly after inoculation with C. higginsianum WT or chbub2Δ (Fig. S8; Table S8). These observations suggest that plant immune response was not induced after chbub2Δ infection.

Discussion

In this study, we demonstrated that BUB2 is an important factor to negatively regulate G1/S progression and initiate septum formation in hemibiotrophic fungi C. higginsianum, M. oryzae, and C. orbiculare. Disruption of BUB2 results in an earlier start of S phase in both C. higginsianum and M. oryzae, and BUB2 disruption mutants had a defect in septum formation governed by SIN and CAR, resulting in lower virulence in host plants. The earlier start of the S phase with cobub2Δ is accompanied by high levels of DNA-replication-associated proteins. The short G1 phase in cobub2Δ causes large changes in the expression of putative virulence-associated genes such as SSPs and CAzymes.

One of the main findings emerging from this study is that the septation phenotype bub2 mutants are different depending on the cell type and fungal species (Fig. 7a). The chbub2Δ did not develop septum or the septum was immature in conidium, but most of them developed septum at the neck of developing appressorium with a larger diameter. In the case of mobub2Δ, the conidium became multiseptate during germination, but the development of septum at the base of appressorium was reduced. In addition, our previous work demonstrated that the C. orbiculare bub2Δ did not have a septation defect either at conidium or at the developing appressorium. These phenotypes were striking because the phenotype of the bub2 mutants of S. pombe and A. nidulans simply leads to a multiseptated phenotype (Simanis, 2003; Kim et al., 2009). It has been well established that septum formation requires the protein of the SIN kinase cascade, associating with CAR assembly and constriction in the fission yeast (Fankhauser & Simanis, 1993; Sparks et al., 1999; Guertin et al., 2000; Hou et al., 2000). The localization of a homologue of SIN component ChDbf2 and the F-actin marker Lifeact in C. higginsianum were well correlated with the septation phenotype of WT and chbub2Δ, suggesting SIN signaling would have a role in committing a signal to the actin network to contract for septum formation similar to the regulation in S. pombe and A. nidulans (Roberts-Galbraith & Gould, 2008; Delgado-Álvarez et al., 2014). In M. oryzae, the future septation site in the neck of the developing appressorium is defined before mitosis by a heteromeric septin ring complex localization, suggesting septin ring formation could be a potential landmark to specify the pre-septation site (Saunders et al., 2010b). In C. orbiculare, the localization of Sep6 to the neck of the developing appressorium was detected more frequently in cobfa1Δ than in WT, suggesting Bub2/Bfa1 complex is involved in the septin localization/degradation, or this might be the secondary effect from the short length of G1 phase in cobfa1Δ. When considered together, we speculate that the Bub2/Bfa1 function in the three hemibiotrophic fungi we analyzed as follows (Fig. S9). The length of the G1 phase is controlled by Bub2/Bfa1 complex, and probably once cells respond to extracellular signals, Bub2/Bfa1 is released and Tem1 is activated to induce signal transduction to express DNA-replication-related genes. Then, potential landmark proteins (e.g. septin) are located to the future septation site. During mitosis, Tem1 is activated again and triggers an SIN signal transduction to relocate Dbf2 to the potential landmark protein. Then, actin is recruited to Dbf2 to form the CAR assembly, leading to initiation of septum formation with chitin synthases activity. In the absence of Bub2, the DNA-replication-related genes are expressed at an improper time by constitutively activated Tem1, leading to earlier transition from the G1 phase to the S phase. Consequently, the short length of the G1 phase causes improper localization of landmark proteins, and Dbf2 becomes unable to relocate to proper future septation sites during mitosis probably because of the mislocalization of landmark proteins. We speculate that Bub2/Bfa1 function for G1 arrest would be analogous in the three hemibiotrophic fungi, but the sequential cell cycle progression and septum formation might be variable, probably depending on their spatial reasons (e.g. cell size, cell shape) or localization of potential septation landmark proteins.

Details are in the caption following the image
The varying phenotype of nuclear division, septum formation, and infection structure in the bub2 mutants. (a) Schematic diagram to show the nuclear division and septum formation. The bub2Δs in three hemibiotrophic fungal pathogens display multinucleation due to the earlier release from G1 arrest and varying septation phenotype during appressorium formation. This contrasts with common patterns observed in Schizosaccharomyces pombe during fission and Aspergillus nidulans during spore germination. Nuclei are represented by blue circles. (b) Model depicting the infection structures in the three hemibiotrophic fungi on the host plants. Differentiation of functional appressorium and the plant immune response are different as a consequence of their defects in cell cycle and septum formation. Plant immune response induced by cobub2Δ is shown in yellow.

In this study, we showed the impaired cell cycle progression and septum formation resulted in avirulent phenotype in C. higginsianum and M. oryzae with more severe defects to differentiate melanized appressorium compared with C. orbiculare bub2Δ (Fig. 7b). The conidia of the chbub2Δ differentiated only one rounder shape from one side of conidium, instead of the lobed appressoria from both side of conidia in WT. In M. oryzae, the deletion of BUB2 affected the differentiation of melanized appressoria, suggesting a more severe phenotype than chbub2Δ, and both bub2 mutants of C. higginsianum and M. oryzae developed fewer infection hyphae than WT did. Since shaping conidia, appressoria, and infection hyphae require proper coordination of apical and polar growth that are well correlated with the activity of cyclin-dependent kinases (Pérez-Martín et al., 2016), these results suggest that bub2 mutants impair the proper switch from apical to polar growth because of the shorter G1 phase duration. Moreover, a septum at the neck of appressorium is likely to be important for functional appressorium formation to separate the appressorium from the germ tube. In M. oryzae, temperature-sensitive mutants of SEP1 enhanced septation and multinucleation, resulting in failure to differentiate mature appressorium (Saunders et al., 2010b). In U. maydis, deletion strains of genes encoding formin Drf1, Cdc42-specific guanine nucleotide exchange factor Don1, or Ste20-like kinase Don3 abolished formation of retraction septa of fungal hyphae and reduced appressorium formation (Freitag et al., 2011). In Colletotrichum sp. and M. oryzae, generating the enormous turgor and melanization are important for penetration of the cuticle and cell wall (de Jong et al., 1997; Thines et al., 2000). Thus, the melanization defect of bub2Δ was presumably correlated with the septation defect at the neck of appressorium. We speculate septum formed at the neck of developing appressorium might be used for the accumulation of massive amounts of glycerol, components related to melanin biosynthesis, or organelles such as peroxisomes so that appressoria become functional to rupture plant cuticle layers.

Unlike the induced plant immune response upon infection with cobub2Δ, there was no induction in A. thaliana after infection with chbub2Δ. The differentially expressed genes encoding CAzymes or SSPs in cobub2Δ were well conserved in C. higginsianum as 76% and 50% conservation for CAzymes and SSPs, respectively. This suggests that their expression patterns in chbub2Δ might not be completely analogous to the one in cobub2Δ as a consequence of more severe defects of cell cycle and septum formation. In addition, the secretion of these proteins might be impaired in chbub2Δ due to the defects in differentiation of melanized appressorium and sequential penetration peg formation (Kleemann et al., 2012; Irieda et al., 2014). At the same time, the plant defense response against the fungal effectors or microbe-associated molecular pattern might also be different between the host plants.

When considered together, our results suggest that infection of plant tissue by Colletotrichum sp. and M. oryzae requires cell-cycle-dependent morphogenetic transitions to enable appressorium formation, thereby enabling host cell penetration. This temporal sequence of morphological transitions by the fungus is tightly coordinated with cell cycle regulation and septum formation controlled by BUB2 to establish plant infection. These observations open the possibility that plant pathogenic fungi use novel functional strategies by regulating cell cycle progression and septation.

Acknowledgements

We are grateful to Hironori Koga in the Laboratory of Plant Protection, Ishikawa Prefectural University for providing M. oryzae Hoku-1 strain, and Yukio Tosa, Hitoshi Nakayashiki, and Kenichi Ikeda of Laboratory of Plant Pathology, Kobe University for providing barley seeds and valuable suggestions. We thank Ken Shirasu, Gan Pamela, and Ayako Tsushima of the RIKEN Center for Sustainable Resource Science for kindly providing genome information of C. orbiculare for the microarray analysis and the nucleotide sequence of ChBUB2. This work was supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science, and Technology (grants 24248009, 15H05780, and KH20140023) and by the Mitsubishi Foundation (no. 24110). We thank Beth E. Hazen for carefully reading the article and giving valuable suggestions.

    Author contributions

    FF conceived and performed experiments. SK performed cytological analysis and virulence assays. TN performed microarray analysis. NK performed cytological analysis. YK supervised the project; FF and YK wrote the manuscript.