Investigating the cell and developmental biology of plant infection by the rice blast fungus Magnaporthe oryzae

Magnaporthe oryzae is the causal agent of rice blast disease, the most widespread and serious disease of cultivated rice. Live cell imaging and quantitative 4D image analysis have provided new insight into the mechanisms by which the fungus infects host cells and spreads rapidly in plant tissue. In this video review article, we apply live cell imaging approaches to understanding the cell and developmental biology of rice blast disease. To gain entry to host plants, M. oryzae develops a specialised infection structure called an appressorium, a unicellular dome-shaped cell which generates enormous turgor, translated into mechanical force to rupture the leaf cuticle. Appressorium development is induced by perception of the hydrophobic leaf surface and nutrient deprivation. Cargo-independent autophagy in the threecelled conidium, controlled by cell cycle regulation, is essential for appressorium morphogenesis. Appressorium maturation involves turgor generation and melanin pigment deposition in the appressorial cell wall. Once a threshold of turgor has been reached, this


Introduction
The blast fungus Magnaporthe oryzae (synonym of Pyricularia oryzae) [1] is able to infect more than 50 different grass species, including staple crops such as rice, millets and barley [2]. It is estimated that rice blast disease causes losses of around 6% of the global rice harvest every year [3] across all rice-growing regions of the world, but epidemics routinely cause up to 30% yield losses [4]. Rice blast therefore represents a severe problem in the 85 countries where rice is grown [5] and to 50% of the world's population who depend on rice as their main source of calories [6].
M. oryzae furthermore has the capacity to jump from one host to another. Wheat blast, for example, first appeared in 1985 in Brazil following a likely host jump from a grass-infecting isolate of the fungus [7]. Increases in global trade since that time, have allowed the disease to spread, emerging in 2016 in Bangladesh, where it now threatens wheat production [8], with potential to spread to India, the world's second largest wheat producer [9]), and most recently to Zambia [10]. When considered together with the ongoing rice blast disease pressures world-wide, M. oryzae represents a very significant threat to global food security. Learning about the basic biology of blast disease is therefore important if new disease-control strategies are to be developed.
The blast fungus has been extensively studied over the past three decades, facilitated by its genetic tractability and development of a suite of molecular genetic tools and genomic resources [11]. Indeed, M. oryzae is now a model system that has revealed important concepts associated with fungal-plant interactions [1,11], such as the molecular basis of appressorium morphogenesis [12,13] appressorium function [4,14], secretory processes associated with effector proteins [15], structure-function relationships governing pathogen recognition by host immune receptors [16,17] and pathogen genome organisation [18,19].
This video article focuses on a series of investigations that have provided new insight into the manner in which the fungus is able to infect rice plants. We start by reviewing the morphogenetic changes associated with development of an appressorium on the rice leaf surface. We highlight cellular changes accompanying appressorium development and how infection-related morphogenesis is controlled by cell cycle progression. We then describe the manner in which M. oryzae re-models its cytoskeleton during appressorium maturation, leading to formation of the rigid penetration peg, which ruptures the leaf cuticle. Next, we outline the biology of plant infection by the rice blast fungus, revealing the extraordinary capacity of this pathogen to proliferate within living host tissue and overwhelm plant defences.
The use of quantitative 4D imaging and high-resolution video microscopy has provided unparalleled insights into the biology of this devastating plant pathogen and allows new studies to take place with far greater resolution than was previously possible.

Cell cycle control and regulated autophagy are necessary for appressorium formation
Conidia of M. oryzae germinate rapidly on the leaf surface, forming a polarised germ tube within two hours of contact with a hydrophobic surface. The spore sticks tightly to the leaf surface by means of spore tip mucilage, released from an apical compartment upon hydration [20]. The germ tube usually emerges from the apical cell of the conidium, extending for a short distance (15-30 µm see Video 1 from 2:00), while adhering tightly to the underlying surface, before swelling and hooking at its tip to form a specialised infection structure, the appressorium ( Figure 1A, B, Video 1 from 2:15 to 3:35) [21]. Appressorium differentiation requires S-phase to have been completed by the nucleus within the germinating conidial cell [22]. A single round of mitosis is then necessary to enable maturation of the appressorium [22,23]. As the appressorium matures, its cell wall becomes lined with a thick layer of melanin necessary for the development of turgor by the appressorium (Video 1, 4:00-7:50). At the same time, glycerol accumulates to molar concentrations to generate hydrostatic turgor due to rapid influx of water into the cell. The conidium undergoes an autophagy-dependent process leading to cell death, trafficking the contents of all three conidial cells into the appressorium (Video 1, conidial cell death is achieved by 10:00), which is necessary for infection [23].
Conidial collapse has been reported to involve ferroptosis as the ultimate mechanism leading to cell death [24]. To investigate nuclear division, a strain of the fungus expressing a Histone H1: green fluorescent protein (GFP) gene fusion [23] was imaged during appressorium differentiation over a period of 24 hours. After the first round of mitosis (Video 2, 4:00-4:10), one daughter nucleus from the germinating conidial cell migrates into the incipient appressorium, leaving the remaining nuclei to be degraded by autophagy-mediated conidial cell death. ( Figure 1C, Figure 2A, B), as shown in Video 2 (conidial nuclei degraded by 13:45) [23,25]. During appressorium maturation, the nucleus in the appressorium then arrests in G1 before progression through S-phase, which is a necessary pre-requisite to re-polarisation [26,27]. Mutants impaired in melanin accumulation do not mature or trigger the S-phase checkpoint. A turgor-dependent cell cycle checkpoint therefore regulates appressorium function, leading to cytoskeletal re-organization and penetration peg emergence ( Figure 2C) [22]. After re-polarisation and plant infection, the appressorium remains mitotically active [28], providing nuclei into nascent invasive hyphae [29].

The Pmk1 MAP kinase signalling pathway regulates appressorium morphogenesis
The development of appressoria in response to the hard, hydrophobic leaf surface and absence of exogenous nutrients also requires the Pmk1 MAP kinase signalling pathway, which is critical to invasive growth by M. oryzae [30], a function conserved in many diverse pathogenic fungal species [31]. Pmk1 is a homologue of the Saccharomyces cerevisiae Fus3 kinase associated with pheromone signalling [32] and the wider pathway links cell surface perception with appressorium morphogenesis [12,31]. As a consequence, pmk1 mutants fail to form appressoria, and do not undergo autophagic conidial cell death, as shown in Figure   2D, E. Pmk1 also, however, plays a role subsequent to appressorium-mediated infection. A conditional, analogue-sensitive mutant of Pmk1 (pmk1 AS ) is, for example, unable to move from cell-to-cell at pit field sites, containing plasmodesmata, during rice tissue invasion when inactivated by the specific kinase inhibitor 1 napthyl-PP1 ( Figure 2F) [32,33]. Pmk1 is activated by a MAPKK (Mst7) and a MAPKKK (Mst11) [34], which in turn are regulated by a putative scaffold protein Mst50 [35]. The Pmk1 MAPK cascade controls the activity of a very large number of downstream targets involved in development and pathogenesis. The transcription factors Hox7 and Znf1, for example, are essential for appressorium formation while Mst12 is required for penetration and invasive growth [36][37][38]. Consequently, the Pmk1 MAPK pathway is fundamental to development of appressoria, although how it is activated by cell surface signals, in particular, is still not well understood.

Septin-dependent re-polarisation of the appressorium
Once formed, appressoria develop enormous turgor of up to 8.0MPa, by developing a melanin-rich cell wall (Video 1) that is able to retain glycerol and other polyols. [39] A critical threshold of turgor allows the fungus to re-orientate its cytoskeleton, and transition from isotropic expansion to anisotropic, polarised, growth during plant infection. A turgor-sensing histidine aspartate kinase, Sln1, is necessary for modulation of turgor, and acts as a regulator of the downstream pathways required for appressorium repolarisation [40]. These include, most notably, the aggregation of a hetero-oligomeric complex of septin GTPases which form a toroidal structure at the base of the appressorium. Septins re-organise F-actin to the precise point of plant infection [41]. A M. oryzae strain expressing Sep5-GFP under control of the native Sep5 promotor allows visualisation of septin ring dynamics during a period of 0-24 hours, when conidia are incubated on hydrophobic glass coverslips (Video 3; Figure 3A).
Septin recruitment to the appressorium pore begins to occur markedly from 7:00 (Video 3) and the ring then becomes apparent by 9:48, showing some constriction by 16:00. The ring is then maintained in appressoria incubated on a non-yielding surface. By contrast, on rice leaf sheath which can be penetrated by appressoria, the ring forms in the same manner but then undergoes further constriction to a diameter of approximately 0.9-1.1m after 28 hpi when the penetration peg is formed ( Figure 3A, Figure 4, Video 3 and Video 4). A recent study using quantitative 4D widefield fluorescence imaging has revealed the spatiotemporal dynamics of F-actin and septin ring recruitment and organisation at the appressorium pore [42]. The septin ring provides cortical rigidification and acts as a diffusion barrier for the action of polarity determinants, endocytic proteins, the exocyst complex, and actin-binding proteins [41,43]. In addition to F-actin, the microtubule cytoskeleton is re-oriented in the direction of cuticle penetration [42]. Organisation of septins in the appressorium involves very long chain fatty acids (VLCFAs), that act as mediators of septin interactions at membrane interfaces. Inhibiting VLCFA biosynthesis therefore prevents rice infection by M. oryzae providing a new potential fungicidal target [44]. Assembly of the septin appressorium pore complex is also controlled by the Nox2/NoxR complex which regulates septin-mediated cytoskeletal dynamics. The actinbinding protein gelsolin, for instance, which regulates actin dynamics by uncapping free barbed ends to promote actin polymerisation, is a likely target for regulation by ROS [45] and an important component of the appressorium pore. We generated a M. oryzae strain expressing Gelsolin-GFP under control of the native Gelsolin promotor. Using 3D maximum projection, Z-stack images were captured at 24hpi ( Figure 3B, Video 5) [45]. Gelsolin forms a highly organised toroidal structure situated at the base of the appressorium (Video 5) in a septin-dependent manner [25]. Septin and F-actin ring formation also requires the Pmk1 MAP kinase pathway and the putative downstream transcription factors, such as Mst12 [41] and Hox7 [37], as shown in Figure 3C. The turgor sensing Sln1 kinase is necessary for determining when a threshold of hydrostatic pressure has been reached in the appressorium to enable repolarisation. As a consequence, Δsln1 mutants are unable to organise septins at the site of penetration and a clear ring structure does not form ( Figure 3B). Sln1 acts in parallel with the protein kinase C cell integrity pathway to suppress melanin biosynthesis and the cAMPdependent signalling protein kinase A pathway, thereby modulating turgor generation [40]. In addition, the turgor-dependent S-phase checkpoint is triggered enabling septin mediated plant infection ( Figure 5) [22,40]. Another significant factor in enabling the appressorium to function is its attachment to the leaf surface. Critically, the force of fungal attachment must exceed the force of penetration at the peg, otherwise the cell would simply lift off the surface. It has been shown that spermine synthase (SPS1) is important for surface attachment by buffering reactive oxygen species generation in the endoplasmic reticulum, thereby facilitating efficient mucilage secretion that provides the tight seal necessary for the appressorium to function. In the absence of Sps1, appressoria do not adhere tightly to the plant surface, impairing infection [46]

The cell biology of invasive growth by M. oryzae
Once the fungus has penetrated the leaf, it develops primary invasive hyphae (IH) that invaginate the plant plasma membrane within the first epidermal cell colonized. The membrane around the invasive hyphae is known as the extra-invasive hyphal membrane (EIHM), but is entirely a plant-derived membrane which tightly surrounds the fungal cell wall [47,48]. The fungus secretes a battery of effectors proteins during this time, which are either directed to the apoplast-the space between the fungal cell wall and the EIHM -or the cytoplasm of the host cell. Fungal effectors suppress plant host immunity to facilitate colonisation of plant tissue ( Figure 6A, B) [47,49]. A membrane-rich structure forms at the tip of the primary IH and the fungus appears to then bud from this point, whereby the membranerich structure forms the biotrophic interfacial complex (BIC), a plant derived membrane-rich body located outside the fungal cell wall [15,50]. Cytoplasmic effectors, such as AVR-Pik and Pwl2, accumulate at the BIC and are also translocated into plant cells. A bright single punctum of AvrPik-GFP is, for example, observed defining the BIC structure, as shown in Figure 6A.
By contrast, apoplastic effectors such as Bas4 outline the invasive hyphae contained within the EIHM ( Figure 6B) when expressed as a Bas4:mCherry fusion protein, although the BIC is also visible (Video 6). Secretion of effectors to these distinct destinations is directed by different secretory pathways. Apoplastic effectors are secreted in a Golgi-dependent, brefeldin A-sensitive manner to the invasive hyphal tip. By contrast, cytoplasmic effectors are secreted from the BIC-adjacent cell into the BIC in an exocyst-dependent pathway that is brefeldin Ainsensitive [15]. It has been shown that the TOR (Target-Of Rapamycin) nutrient-signaling pathway is important in mediating membrane integrity at the plant-fungal biotrophic interface.
Deletion of a novel vacuolar protein gene IMP1, led to mutants impaired in BIC formation which released apoplastic effectors into the plant host cytoplasm [51]. The deployment of a large battery of effector proteins enables M. oryzae to overcome plant immunity and rapidly move from the first invaded cell to neighbouring cells, through pit fields. Interestingly, this involves a specific morphogenetic transition in which the hyphal tip swells when it makes contact with a pit field, followed by severe hyphal constriction to around 360 nm in diameter (equivalent to a pit field) and then emergence of a new invasive hypha in the neighbouring cell. This morphological transition is regulated by the Pmk1 MAP kinase [33,47] (Figure 6C, and Video 7). When an analogue-sensitive pmk1 AS mutant is allowed to infect a host cell and then treated with the 1NA-PP1 inhibitor, it becomes trapped within the epidermal cell [15]. The Pmk1 MAPK pathway is therefore required for septin-dependent constriction of invasive hyphae at cell wall crossing points in the same way as it is necessary for appressorium morphogenesis. Each crossing point can be clearly seen as being preceded by pronounced swelling of the invasive hypha (Video 7, 517 min), followed by severe constriction as the hypha moves into the next cell (Video 7, 641-713 min). Often mitosis is seen to occur just after the time of cell-to-cell movement [29,52], which can be visualised by observing tissue invasion in a M. oryzae strain expressing H1-GFP (see Video 8, from 34 min, in which the arrow shows cell crossing point, with mitosis occurring shortly afterwards by 36 min). Appressorium formation and cell wall crossing therefore have many common features requiring swelling into a yeast-like hemispherical cell, followed by re-polarisation and emergence of a narrow penetration peg or IH peg, respectively. The cell wall crossing structure has been termed the 'transpressorium', because of the resemblance to appressorium formation (see Video 9 in which the terminal swelling and infection peg formation can be seen in the 3D image.) [53].
The regulation of transpressorium formation by Pmk1 and the requirement for septin-mediated invasion [15] are consistent with the morphological conservation.
The suppression of immunity responses by M. oryzae enables the fungus to colonise leaf tissue very efficiently. This is illustrated by Video 10 in which a strain expressing a nucleolar marker Fib1-GFP is seen to rapidly colonise plant tissue between 24 and 38h after inoculation. During this period the number of fungal nuclei observed increases from 18 to 85 in the field of view shown, highlighting the speed of cell division that accompanies fungal invasive growth in this 14h period. A three-dimensional reconstruction of invasive hyphae demonstrates the extent of host colonization, as shown in Video 11. Here, fungal invasive hyphae are visualised using fluorescent wheat germ agglutinin, within propidium iodidelabelled plant tissue. Invasive hyphae are large bulbous and branched as they spread within epidermal and mesophyll cells, but more elongated as they move into vascular tissue. The colonisation of tissue is extremely rapid, with plant cells losing viability at the centre of a fungal colony as the fungus continually moves into viable cells, maintaining EIHM integrity as it does so. The central part of the colonies, which becomes visible as a disease lesion, will then produce aerial hyphae that develop into conidiophores bearing sympodial arrays of conidia.

Conclusions
In summary, live cell imaging studies have dramatically expanded our understanding of how rice plants are infected by the blast fungus M. oryzae [22,40,41,43,45]. Generation of functional appressoria provides the gateway for establishing rice blast disease, and it is now clear that this is a highly orchestrated developmental process, requiring several important prerequisites-such as perception of the hard-hydrophobic leaf surface, and response to cutin monomers or components of plant epicuticular waxes [21]. Next, the fungus utilises two key signalling pathways, the cAMP protein kinase A pathway and the highly conserved Pmk1 MAP kinase signalling pathway, regulated by G-protein signalling in the developing germ tube tip.
The Pmk1 signalling pathway regulates appressorium formation, maturation, invasive growth and infection of plant tissue [32]. Pmk1, for instance, regulates expression of many genes encoding secreted fungal effectors proteins implicated in the suppression of host immunity, in addition to its better known role in symmetry-breaking and morphogenesis [33]. Appressorium development is, however, also tightly coupled to cell cycle progression [4,22,26], which is itself linked to nutrient availability and the critical role of TOR kinase as a growth regulator that is fundamental to the control of autophagy [14,54,55]. Once formed, the appressorium undergoes further changes and the role of glycerol and melanin biosynthesis in turgor generation are well established, but still lacking in specific details, particularly regarding gene regulation and the precise enzymatic activities necessary at each step in these pathways. A turgor-sensing mechanism has also now been reported for appressoria, linking turgor control to re-polarisation of the infection cell.
Many questions, however, remain to be answered regarding operation of an appressorium. How does the turgor-sensing complex work in M. oryzae and what are its components? How precisely does Sln1 negatively regulate melanin biosynthesis, interact with the Pkc1 cell integrity pathway and negatively regulate glycerol production via the cAMP/PKA pathway? What is the precise trigger for septin aggregation at the appressorium pore and how does this differ from septin recruitment in well-studied processes, such as bud formation in S.
Following leaf infection, the biology of tissue colonisation by M. oryzae and the highly effective strategies deployed by the fungus to suppress host immunity, both metabolic and effector-driven, are becoming clearer [57][58][59]. The developmental transitions involved in traversing cell junctions within a rice leaf are just as striking and developmentally complex, as those on the leaf surface. The transpressorium has significant developmental parallels to the appressorium, having evolved to traverse the same type of structural barrier. Many components are conserved, such as the Pmk1 MAP pathway and septins, for example.
However, there are clear differences, such as need to move from one viable cell to another, all the time maintaining EIHM integrity [33,47] and the requirement for host immunity to be suppressed at plasmodesmata, where such responses are often focused [60]. Critical questions include, how does the fungus identify plasmodesmata-rich pit fields as potential crossing points and suppress host immune response at these cell junctions? How are effectors secreted by invasive hyphae and delivered across the plant plasma? And what is the precise role of each effector protein and why is such a large arsenal of effectors necessary to establish blast disease [49].   A. Life cycle of M. oryzae. Infection begins when a three-celled conidium lands and attaches to the hydrophobic surface of a rice leaf. The spore germinates producing a long narrow germ tube that differentiates into an appressorium. The single celled appressorium matures, and the three celled conidium collapses and dies in a programmed process requiring autophagy and ferroptosis. The appressorium melanises and generates enormous cellular turgor pressure. This is translated into mechanical force leading to rupture of the rice leaf cuticle.    Images were captured after 28hpi. Scale bar =10 µm.  were inoculated onto rice leaf sheath and images were captured every 10 minutes from the point of penetration (24hpi) for 16 hours. Scale bar =20 µm.

Video 1. Live cell imaging of appressorium development and maturation.
Conidia were harvested from wild type strain Guy11 and inoculated onto glass coverslips. The movie was captured using a Leica SP8 laser confocal microscope 2-15hpi. The movie is a maximum projection of Z-stack, frames were taken every 5 min and are displayed at 15 frames per sec. Time scale is in hours: mins. Scale bar =10 µm.
images were captured at 28hpi using a Leica SP8 laser confocal microscope. Scale bar= 10µm.

Video 7. Tissue colonisation by M. oryzae.
Conidia were harvested from wild strain Guy11 and inoculated onto rice leaf sheath. Invasive growth was imaged every 10 minutes for a duration of 16 hours using a Leica SP8 laser confocal microscope. Arrows indicate the crossing points of invasive hyphae into neighbouring cells. The movie is a maximum projected Z-stack, frames were taken every 10 min and are displayed at 7 frames per sec. Time scale is in mins. Scale bar= 20µm.

Video 8. Nuclear division during invasive growth by M. oryzae
Rice leaf sheath tissue inoculated with Guy11 expressing H1-GFP showing nuclear division during cell-to-cell movement by the fungus. Arrow indicates cell crossing point where transpressorium forms, which is shortly followed by mitosis and nuclear movement to the adjacent cell where it appears close to junction with the next cell (arrowed). The movie is a maximum projected Z-stack, with images recorded every 10 min, displayed at 15 frames per sec. Scale bar = 10 µm oryzae Guy11 within infected rice tissue stained with propidium iodide. Scale bar = 50µm.