Background

Communication is a ubiquitous primitive characteristic developed by all living species. The ability to communicate effectively may affect mating, predation, competition, dominance hierarchy, signal modalities, and survival [1,2,3]. This complex mechanism starts when a given organism (the sender) secretes in the environment a self-produced molecular signal (the message) that alters the behavior of another organism (the receiver) [1, 3]. Communication also happens at the cellular level. This so-called cell-to-cell communication creates a complex signaling network that involves different extracellular signals and distinct cell types that regulate several pathways [4,5,6]. Inter- and intraspecies cell-to-cell communication has been widely studied in fungi to address biological functions including the secretion of pheromones to attract the opposite sexual partner, the production of quorum sensing molecules controlling the expression of virulence factors or morphological changes, and the regulation of cell fusions during vegetative growth [6,7,8,9].

The fungal mycelium is formed by three integrated processes, including hyphal extension, branching, and vegetative hyphal fusion (VHF) (also known as anastomosis) [10]. In this last-mentioned process, two growing cells with identical vegetative compatibility loci engage in cell-to-cell communication, which is thought to involve the secretion of unknown diffusible molecules and results in re-direction of polarized hyphal growth toward each other. After physical contact, the cell walls are remodeled, the plasma membranes fuse, and the two interconnected cells exchange cytoplasm and organelles [11]. When the anastomosed individuals are vegetatively incompatible, the two fused cells rapidly collapse following DNA degradation by programmed cell death, or they are severely inhibited in their growth [12]. It is widely accepted that mycelial network formed through VHF facilitates the intra-hyphal communication, translocation of water and nutrients, and signal molecules, which improve the general homeostasis and spatial expansion of the fungal colony [13, 14]. In some fungi, hyphal fusion is required for pathogenicity and host adhesion [15,16,17]. Cell fusions can also occur between germinating conidia, the asexual spores of many fungi. The process of fusion between germinating conidia involves the formation and interaction of specialized hyphae, so-called conidial anastomosis tubes (CATs). CATs are thinner and shorter than vegetative hyphal fusions (VHFs), and its induction is dependent on nutrient deprivation and initial cell density [18]. Cell fusions may serve to improve colony establishment, as well as to increase the genetic variability by facilitating heterokaryosis and parasexual recombination [18]. Gene or chromosome transfers by cell fusions between individuals of the same or different species allow certain fungi to acquire pathogenicity or to broaden their host specificity [19,20,21]. Albeit non-self-anastomoses are described [22, 23], this might be a rare event in nature.

In the last decades, different studies about the molecular mechanisms underlying cell fusion identified several mutants defective in anastomosis, revealing that fungal communication and fusion are complex mechanisms that encompass several signaling pathways [6]. The best-characterized mutant is the soft (So) gene of Neurospora crassa [24]. So, it is proposed to be the scaffold protein for some mitogen-activated protein kinase (MAPK) from the cell wall integrity (CWI) signaling pathway implicated in the regulation of different fungal processes [25,26,27]. For instance, So contributes to septal plugging during hyphal injury or damage caused by environmental stresses [28, 29]. Nevertheless, So gene has an essential role in the hyphal anastomosis, presumably by regulating the secretion or perception of an undefined chemoattractant in an oscillatory manner with the Fus3 (orthologous to MAK-2 in N. crassa) from the MAPK pheromone response pathway, as demonstrated for N. crassa [11]. Beyond N. crassa, the So was also characterized in other model organisms, plant pathogens, and endophytic fungi [15,16,17, 29, 30]. Though all So mutants fail to undergo hyphal fusions, the distinct effects on pathogenicity reported for those null mutants suggest that the biological contribution of anastomosis might depend on the infection strategies developed by different fungal pathogen species.

Zymoseptoria tritici is an apoplastic pathogen with a hemibiotrophic lifestyle and considered the most damaging pathogen of wheat in Europe [31]. This fungus has the ability to undergo morphological transitions in response to the environment, switching between hyphal growth and yeast-like growth [32,33,34]. Hyphae formed from either germinated ascospores (sexual spore), pycnidiospores (asexual spores), or blastospores (asexual yeast-like spores produced by budding) are essential for penetrating wheat leaves through the stomata and colonization of the apoplastic space. After a long asymptomatic phase (which varies depending on the wheat genotype and fungal strain combination) [35,36,37], the onset of the necrotrophic phase is followed by the appearance of lesions, disintegration of host tissue, and formation of asexual fruiting bodies. Though Z. tritici is among the top 10 most studied phytopathogens [38], there is little known about vegetative cell fusion in this organism. To date, it was shown that the deletion of the β-subunit of the heterotrimeric G protein MgGpb1 or ZtWor1, a transcriptional regulator of genes located downstream of the cyclic adenosine monophosphate (cAMP) pathway, results in germ tubes that undergo extensive anastomosis [39, 40].

In this study, we aimed to determine whether vegetative cell fusions play essential biological roles in the lifestyle of Z. tritici. We showed that the ubiquitous ability of Z. tritici to undergo self-fusion was disrupted by the deletion of ZtSof1, affecting both hyphal and germling fusions. The characterization of mutants lacking the MAPK-encoding ZtSlt2 or ZtFus3 indicates a conserved role of the CWI and pheromone response pathways on fungal anastomosis. We found that ZtSof1 contributes to vegetative growth and is required for melanization, but not to maintain the fungal cellular integrity. We discovered that anastomoses are dispensable for pathogenicity, but they are essential for asexual fruiting body development. In the absence of cell fusions, Z. tritici does not undergo asexual reproduction. These findings illustrate the impact of ZtSof1 for fungal development and the importance of vegetative cell fusions for fungal fitness.

Results

Cell fusions in Z. tritici allow the bidirectional transfer of cytoplasmic content but do not enable multinucleated cell formation

We co-inoculated either blastospores or pycnidiospores of both 1E4GFP and 1E4mCh fluorescent strains onto water agar (WA—1% agar in water), a hyphal-inducing medium, to investigate the ability of Z. tritici to undergo self-fusions. Though Z. tritici produces blastospores and pycnidiospores as asexual spores instead of conidium, we used the CAT terminology to define the fusion between germinating spores. CATs formed between blastospores or pycnidiospores germlings happened at high cell density (107 blastospores/mL), starting after 4 h of incubation, but they were frequently observed after 17 h of incubation, resulting in cells co-expressing both fluorescent proteins (Fig. 1b). On the other hand, vegetative hyphal fusions (VHFs) from germinated blastospores or pycnidiospores were noticed at 40 h after incubation (hai) only at low cell density (106 blastospores/mL) (Fig. 1c, d). Multiple interconnections via fusion bridges were observed in all tested morphotypes (Fig. 1b–d). The co-infection of wheat plants using either blastospores or pycnidiospores of 1E4GFP and 1E4mCh strains also resulted in VHFs on the wheat leaf surface (Additional file 1: Fig. S1). Self-fusions and cytoplasmic mixing occurred in the first 48 hai (Additional file 1a-b: Fig. S1a-b). To monitor for nucleus movement enabling multinucleated cell formation, we used the IPO323 ZtHis1-ZtGFP strain [41]. As previously reported, only one nucleus per cell was observed in this strain. No multinucleated septal compartments were observed between the two interconnected hyphae at the fusion bridges or outlying of the fusion points (Additional file 2: Fig. S2). These findings suggest that cell fusions culminate in cytoplasmic mixing in Z. tritici, but it does not lead to the generation of multinucleated cells in this uninucleated fungus.

Fig. 1
figure 1

Vegetative cell fusions of Zymoseptoria tritici. a Blastospores of a GFP- or mCh-tagged 1E4 strain appeared only in a one-color channel of the fluorescence microscope. b High initial cell density (1 × 107 blastospores/mL) induced conidial anastomosis tubes (CATs) that were frequently observed after 17 h of incubation. c, d Vegetative hyphal fusions (VHFs) from germinating blastospores or pycnidiospores of a GFP- or mCh-tagged 1E4 strain were induced at lower initial cell density (1 × 106 blastospores/mL) and were noticed after 40 h of incubation. Cell fusions may result in the continuous mixing of cytoplasmic content of individuals expressing GFP or mCh fluorescent proteins. Black arrows or white triangles point to the CATs or VHFs, respectively. White asterisks point to the fusion bridges formed between individuals expressing identical fluorescent proteins

Full size image

The mutual perception or response to genetically identical fusion partners requires the ZtSof1 gene and different MAPK pathways

So has an essential role in self-anastomosis [24]. To determine whether this gene plays the same role in Z. tritici, we identified the So orthologous (ZtSof1) in the Z. tritici genome (Mycgr3G74194 or Zt09_7_00503), which consists of a 3794-bp open reading frame that encodes a polypeptide of 1227 amino acid and is widely distributed within the Dothideomycetes (Additional file 3: Fig. S3). We phenotyped the ΔZtKu70, ΔZtSof1, and ΔZtSof1-comp mutants generated on the 1E4 strain genetic background for the presence of interconnected individuals through anastomoses. Fusion bridges between blastospore germlings and filamentous hyphae were only observed for those strains possessing the ZtSof1 gene, but none between ΔZtSof1 mutant cells (Fig. 2a and b), demonstrating that ZtSof1 plays an essential role in anastomosis between genetically identical Z. tritici strains.

Fig. 2
figure 2

ZtSof1 is required for vegetative cell fusion in Zymoseptoria tritici. All mutant lines derived from the 1E4 strain. a Fusion bridges between blastospore germlings were observed at high cell density for ΔZtKu70 and ΔZtSof1-comp strains but not in the ΔZtSof1 mutant. Black arrows indicate CATs. b VHFs were noticed at low cell density and only for those strain possessing the ΔZtSof1 gene. White triangles point to the fusion bridges between two fused hyphal cells

Full size image

To ensure the failure of cytoplasmic exchange on those individuals lacking the ZtSof1 gene, we mixed blastospores of each tested strain with 1E4GFP blastospores in a 1:1 ratio and traced the spores up to 40 h. Hyphal fusions and the continuous streaming of cytoplasmic green fluorescence coming from the fusion with the 1E4GFP strain were observed for all tested combinations (Additional file 4: Fig. S4, and Additional file 5: Fig. S5), except for the combination of ΔZtSof1 and 1E4GFP spores (Additional file 6: Fig. S6). These findings confirmed that ZtSof1 is required in both fusion partners for a mutual recognition previous to the anastomosis.

To determine whether the CWI and pheromone response pathways also contribute to anastomoses in Z. tritici, we incubated the knocked-out ZtSlt2 (orthologous to MAK-1 in N. crassa) and ZtFus3 (orthologous to MAK-2 in N. crassa) mutants on WA plates. No hyphal fusion was observed for both mutants. Unlikely, fusion bridges were regularly found between IPO323 wild-type or ΔZtSlt2-complemented fungal cells (Additional file 7: Fig. S7). These results indicate the recruitment of both CWI and pheromone response MAPK signaling cascades for the regulation of self-fusion in Z. tritici, as it has been observed in N. crassa [11].

Deletion of ZtSof1 affects growth and melanization

To assess whether vegetative cell fusions affect fungal development, we determined the radial growth of ΔZtKu70, ΔZtSof1, and ΔZtSof1-comp on different nutritional environments inducing different morphotypes. We observed a slightly reduced growth of ΔZtSof1 colonies when grown on WA plates, a condition that induces hyphal growth, compared to those possessing the ZtSof1 gene (Additional file 8a: Fig. S8a). On average, the colony radii were 5.31 ± 0.12 for ΔZtKu70, 4.87 ± 0.10 for ΔZtSof1, and 5.50 ± 0.09 for ΔZtSof1-comp [radial growth (mm) ± standard error] (Additional file 8c: Fig. S8c). Nevertheless, ΔZtKu70 and ΔZtSof1-comp generated colonies with highly hyphal dense margins, while ΔZtSof1 exhibited sparse filamentations at the colony periphery (Additional file 8b: Fig. S8b). Unlikely, the ΔZtSof1 mutant grew significantly faster than the ΔZtKu70 and ΔZtSof1-comp on the nutrient-rich PDA medium, a condition that induces blastosporulation (yeast-like growth) (Additional file 8e: Fig. S8e). Over time, the relative growth rate of ΔZtKu70 and ΔZtSof1-comp were respectively 32% and 16% lower than the ΔZtSof1 mutant. No morphological differences were detected between blastospores produced by the tested strains (Additional file 8d: Fig. S8d). The data suggest that ZtSof1 mutation may affect fungal growth in a morphotype-dependent manner. Consistent with these findings, we found that the deletion of ΔZtSof1 increased blastosporulation in a nutrient-rich liquid medium. At 48 and 72 hai, the ΔZtSof1 mutant produced a significantly higher amount of blastospores than ΔZtKu70 and ΔZtSof1-comp strains (Additional file 9: Fig. S9).

Interestingly, no melanin accumulation was observed in ΔZtSof1 mutant colonies (Fig. 3), demonstrating the impact of ZtSof1 deletion on Z. tritici pigmentation. We postulated that the ΔZtSof1 mutant could either display cellular integrity defects and/or being susceptible to environmental stresses. We tested nine different abiotic stressors, such as temperature, oxidative, osmotic, cell wall, and cell membrane stresses. Overall, no variability in stress responses was noticed among the strains (Fig. 4). However, ΔZtSof1 formed slightly bigger colonies than those from ΔZtKu70 and ΔZtSof1-comp, corroborating the increased growth rate observed for this mutant on nutrient-rich medium (Additional file 8e: Fig. S8e). Therefore, we found no evidence that the ZtSof1 gene is involved in the maintenance of the cell wall integrity of Z. tritici.

Fig. 3
figure 3

Disruption of ZtSof1 impacts the melanization of Zymoseptoria tritici. The defective fusion mutant was significantly less melanized than the ΔZtKu70 and ΔZtSof1-comp strains, which exhibited higher melanin accumulation over time. Bars represent standard errors of the mean gray values on at least 40 colonies. Different letters on the top of the bars indicate a significant difference among the tested strains according to the analysis of variance (ANOVA). The notch displays a 95% confidence interval of the median. Open circles represent the outlier values of each strain. Pictures shown below the bar plot represent the melanization level of ΔZtKu70, ΔZtSof1, and ΔZtSof1-comp strains. Gray value scale (0 = black and 255 = white) is shown on the left

Full size image
Fig. 4
figure 4

ZtSof1 is dispensable for cellular integrity. A serial dilution of blastospore suspensions of ΔZtKu70, ΔZtSof1, and ΔZtSof1-comp strains were exposed for 5 days to nine different stress conditions, including different temperatures (18 °C and 27 °C), oxidative stress (0.5 and 1 mM of hydrogen peroxide (H2O2)), osmotic stresses (1 M sodium chloride (NaCl) and 1 M sorbitol), cell wall stresses (2 mg/mL Congo red and 10 μg/mL Calcofluor white (CFW)), and plasma membrane stress (0.01% sodium dodecyl sulfate (SDS)). The tested strains do not vary on their tolerance to different cellular stressors

Full size image

Hyphal fusions are essential for the development of asexual fruiting bodies

We inoculated a susceptible wheat cultivar with the tested Z. tritici strains to assess the biological role of vegetative cell fusion during the pathogen lifecycle in planta. Typical symptoms caused by Z. tritici infections were visible after 11 days post-infection (dpi), and the disease progression was similar among the plants inoculated with ΔZtKu70, ΔZtSof1, or ΔZtSof1-comp strains (Fig. 5a and Additional file 10: Fig. S10). This result indicates that ZtSof1 is neither required for host penetration nor for the asymptomatic or necrotrophic phases of the fungus. The asexual fruiting bodies (pycnidia) were visible on plants inoculated with those strains possessing the ZtSof1 gene at 14 days post-inoculation (dpi). In contrast, plants infected with ΔZtSof1 never developed pycnidia. The failure to undergo asexual reproduction was observed in a broad range of wheat genotypes (Additional file 11: Fig. S11).

Fig. 5
figure 5

The role of vegetative hyphal fusions for disease progression and pycnidial development. a Susceptible wheat cultivar Drifter inoculated with ΔZtKu70, ΔZtSof1, or ΔZtSof1-comp strains were evaluated up to 21 days post-infection (dpi). All tested strains exhibited similar disease progression, including the onset of the necrotrophic phase at 11 dpi. b Pycnidia production on wheat extract agar. After 20 days of incubation, ΔZtKu70 and ΔZtSof1-comp strains produced brown pycnidium-like structures exuding a whitish liquid similar to the oozed cirrhus-containing pycnidiospores spores observed for Z. tritici-infected wheat plants. In contrast, ΔZtSof1 mutant formed mycelial knots, but those structures never developed in mature asexual fruiting bodies

Full size image

To distinguish whether pycnidium formation would be a consequence of the lack of hyphal fusions or susceptibility of ΔZtSof1 cells to defense compounds produced by the plant, we used a wheat extract agar medium to induce pycnidium formation in vitro. ΔZtSof1 aggregated into mycelial knots, which is the initial developmental stage of the pycnidium, but no asexual reproductive structures were further developed. The pycnidium-like structures formed by ΔZtKu70 or ΔZtSof1-comp strains were exuding a whitish liquid containing pycnidiospores (Fig. 5b). Next, we evaluated the pycnidium formation in planta. The wheat plants infected with 1E4GFP or 1E4GFPΔZtSof1 strains were monitored using confocal microscopy up to 12 dpi. We observed mainly host penetration, initial intercellular hyphal extension, and sub-stomatal colonization at the earlier stages of the plant infection (Additional file 12: Fig. S12 – 6 and 7 dpi). No difference was noticed in fungal development; however, hyphal fusions established during epiphytic host colonization were only detected for 1E4GFP strain. At 8 and 9 dpi, we observed that the primary intercellular hyphae surrounding the stomatal guard cells produced specialized knots from where secondary hyphae emerged and elongated (Fig. 6 and Additional file 12: Fig. S12). These secondary hyphae fused with other nearby hyphae, creating an interconnected network in the sub-stomatal cavity (e.g., for the 1E4GFP strain) or they kept extending as individual hyphae (e.g., for the 1E4GFPΔZtSof1 mutant) (Fig. 6 and Additional file 12: Fig. S12). The combination of sub-stomatal hyphal accumulation and anastomoses generates the mature pycnidium, which supports the asexual reproduction of Z. tritici. On the other hand, the lack of anastomosis stopped the symphogenous development by affecting the interweaving of hyphal branches through hyphal fusion, a crucial mechanism for the formation of pycnidia, and hence, impairing the asexual cycle of the fungus. Therefore, we concluded that hyphal fusions are decisive for the pycnidial development, and the disturbance of this mechanism ceases the asexual reproduction of Z. tritici.

Fig. 6
figure 6

Confocal microscopy images and schematic demonstration of pycnidial development during plant infection. Susceptible wheat cultivar Drifter inoculated with the fluorescent 1E4GFP (wild-type, left) or ΔZtSof1GFP (right) strains were monitored by confocal microscopy at different days post-infection (dpi). For 1E4GFP, the primary intracellular hyphae surrounding the stomatal guard cells produced specialized knots from where secondary hyphae emerge and elongate. After 9 days of infection, these secondary hyphae fuse with another nearby hypha (represented by black circles), creating an interconnected network in the sub-stomatal cavity. The combination of sub-stomatal hyphal accumulation and anastomoses generates the pre-pycnidium at 12 days, which later supports the asexual reproduction of Z. tritici. For the ΔZtSof1GFP mutant, the filamentous hyphae kept extending as individual hyphae, and no fusion points were observed. The lack of anastomosis stops the developmental process of the pycnidium formation. For earlier time points (6, 7, and 8 dpi), please see Additional file 12: Fig. S12

Full size image

Discussion

The evolution of the fungal language allowed a fine-tune coordination of signal senders and receivers driving their ecological diversifications and, consequently, the regulation of complex signaling networks. Here, we explored the functional relationship of a gene involved in cell-to-cell communication and its biological contribution to the development and fitness of a fungal plant pathogen.

Vegetative cell fusion is one of the most important developmental processes of a mycelial fungal colony [10, 42]. The cytoplasmic continuity generated by cell fusion provides adaptative advantages to the interconnected mycelial network by allowing resource sharing and introgression of genetic material [19, 42, 43]. CATs forming at earlier stages of the vegetative growth have been reported in several filamentous fungi [18]. We observed that these specialized fusion bridges require a specific cell density, indicating that Z. tritici may induce CATs at a critical concentration of a self-produced molecule, presumably a quorum-sensing molecule. For instance, Fusarium oxysporum secretes and senses both a- and α-pheromone, and their perception regulates spore germination in a cell density-dependent manner [44]. On the other hand, the induction of VHFs was often observed at lower cell density, indicating that VHFs may be induced at low self-produced molecule concentration or some other environmental signal. Z. tritici is a pleomorphic fungus changing growth morphology according to the environment. Nutrient-limited conditions induce both hyphal growth and fusions. Plant pathogens typically experience nutrient limitations while growing on leaves, and the perception of a nutrient-limited environment may act as a stimulus to induce interconnected mycelial formation in foliar plant pathogens, as observed for the causal agent of anthracnose disease, Colletotrichum lindemuthianum [45]. Both germinating blastospores or pycnidiospores underwent CATs and VHFs in vitro and in planta, supporting our previous observation [34]. Therefore, we demonstrated that vegetative cell fusions are a ubiquitous cellular process, which may have an essential contribution to the lifestyle of Z. tritici.

Nucleus transfer between two encountering hyphae occurs in some fungal species [46, 47]. The consequences of genetic exchange include the formation of viable heterokaryons and the risk of introgression of pathogenic elements or virulence genes [20, 48, 49]. Unlike the majority of fungal species that form a multinucleated hyphal network [47], Z. tritici has only one nucleus per septal compartment [41], which raises the question of whether anastomosis would enable multinucleated cell formation in Z. tritici. We showed that the new cell compartment is occupied by a migrating nucleus coming from a neighboring nucleus divided by mitosis after the fusion bridge between two identical Z. tritici cells. However, cells containing multiple nuclei were not observed neither near nor far from the anastomosis point, indicating that Z. tritici may have evolved to limit the spread of genetic elements and restrict the formation of multinucleated cells. Hence, we cannot discard that multinucleated, homokaryotic, or heterokaryotic cells can be formed at low frequency in nature. These properties could impact the evolution of the pathogen, and therefore, additional studies will be needed to underlie the roles played by the genetic exchange during vegetative cell fusions in Z. tritici.

In the last decades, the identification of fusion-defective mutants has contributed to the understanding of the molecular mechanisms underlying cell communication and fusion, especially by unveiling the interplay of the mitogen-activated protein kinase (MAPK) pathways during chemotropic interactions [11]. MAPK pathways are involved in extracellular signal perception and regulation of diverse genes essential for mating, filamentation, pathogenicity, cell integrity, and stress responses [50,51,52,53,54,55]. Therefore, the deletion of MAPK-encoding genes results in pleiotropic phenotypes due to their involvement in multiple biological processes. Though fungal communication and fusion require the regulation of several genes [6], the cross-talk between cell wall integrity (CWI) and pheromone response, two conserved MAPK signaling pathways, is essential to produce, secrete, and sense the chemoattractant molecule produced during cell fusion [11]. The deletion of Slt2 or Fus3 orthologous genes disrupts the signaling cascade affecting self-anastomosis in filamentous fungi. We used the ZtSlt2 and ZtFus3 to demonstrate the functional conservation of these MAPK pathways in Z. tritici. ZtSlt2 and ZtFus3 are known as essential genes for the pathogenicity and developmental processes of Z. tritici, including vegetative growth, melanization, and pycnidium formation [56, 57]. Here, we showed that both ΔZtSlt2 and ΔZtFus3 mutants were unable to undergo anastomosis. Our results indicate that cell fusions in Z. tritici follow the same signaling mechanism described for N. crassa [11, 58], where the signal sending and receiving are coordinated by genes associated with the CWI and pheromone response pathways. Besides, this is the first report of those pathways regulating cell communication in Z. tritici.

To evaluate the impact of vegetative cell fusion in the biology of a hemibiotrophic fungus, we used the ZtSof1 orthologous of the N. crassa So gene. Contrary to the MAPK-related genes, the characterization of So orthologs results in low pleiotropy [15, 16, 24]. Filamentous fungi lacking So gene were impaired in self-anastomosis [15,16,17, 24], including Z. tritici, in which the deletion of ZtSof1 abolished vegetative cell fusion. We characterized the impact of the fusion deficiency at different developmental stages of Z. tritici. For instance, the ΔZtSof1 mutant exhibited a more asymmetrical hyphal extension in a nutrient-poor medium inducing hyphal growth than the fusion-competent individuals. It was demonstrated that the direction of the nutrient distribution occurs from the central part of the mycelium to outwards, and its streaming speed is driven by anastomosis [42]. Thereby, the reduced colony extension observed for the ΔZtSof1 may be a consequence of the irregular distribution of cytoplasmic content throughout the mycelial colony. In contrast, the induced blastosporulation and larger yeast-like colonies exhibited by the fusion-defective mutant in nutrient-rich environments may be a consequence of the disruption of the CWI pathway or its cross-talk with other MAPK pathways involved in fungal growth. The genetic relationship between vegetative growth and the ZtSof1 gene remains to be elucidated.

We showed that the ΔZtSof1 colonies do not accumulate melanin, resulting in whiter and larger colonies than those formed by ΔZtKu70 and ΔZtSof1-comp strains. These findings suggest a possible trade-off between energy cost for pigment production and growth. The reduction of fungal growth caused by the higher accumulation of melanin was reported before for Z. tritici [59]. Melanins are dark-pigmented secondary metabolites often associated with the fungal cell walls. Though fungi can produce different kinds of melanins, it is suggested that melanization of Z. tritici is only controlled by the polyketide synthase (PKS) gene cluster containing catalytic enzymes and transcription regulators of the 1,8-dihydroxynaphthalene (DHN) melanin [59, 60]. It was demonstrated for plant pathogenic fungi that the deletion of CWI-associated genes inhibited pigmentation by reducing the expression of DHN melanin biosynthetic genes [61,62,63,64]. Therefore, the deletion of ZtSof1 may impair the CWI-regulatory cascade and, consequently, the regulation of DHN-melanin production, resulting in the lack of pigmentation observed for both ΔZtSof1 and ΔZtSlt2 [56] mutants. On the other hand, the CWI, high-osmolarity glycerol (HOG), cyclic adenosine monophosphate (cAMP), and the pheromone response pathways can interact regulating melanization in a cooperative manner [65,66,67]. Consistent with the findings in other fungi, mutants in these signaling pathways, including ΔZtFus3, display altered pigmentation in Z. tritici [33, 39, 56, 57, 68, 69]. Additionally, a QTL mapping study identified ZtSlt2, ZtHog1, ZtGpa1, and ZtFus3, contributing to the melanization of this fungus [60]. Thereby, we cannot discard that the deletion of ZtSof1 may affect the cross-talk between CWI with other signaling cascades impairing fungal melanization. Further investigations are needed to elucidate both hypotheses.

Melanin is postulated to contribute to fungal protection against fungicide and environmental stresses in Z. tritici [59, 60], and its regulation depends on environmental cues and colony development [60], though it is not yet a fully understood mechanism. We used nine different cellular stressors to evaluate whether (i) the defect in melanin accumulation or (ii) the deletion of ZtSof1, the scaffold protein for the MAPK genes from the CWI pathway, affects pathogen stress tolerance. Since the non-melanized ΔZtSof1 mutant displayed the same degree of stress sensitivity than ΔZtKu70 and ΔZtSof1-comp strains, we concluded that ZtSof1 does not act as a scaffold protein for all CWI pathway functions. Consistent with our findings, the model fungus Sordaria macrospora, the PRO40 (orthologous to ZtSof1), operates as a scaffold for the CWI-encoding genes during fungal development, hyphal fusion, and stress response, but not for growth under cell wall stress agents [25]. Our results showed that melanin accumulation of Z. tritici does not benefit fungal survival in harsh environments, at least for the stressful conditions tested in this study.

We demonstrated that VHFs are dispensable for the pathogenicity of Z. tritici. The fusion-defective ΔZtSof1 mutant displayed a similar host damage progression than those individuals possessing the gene. This finding exemplifies the distinct effects of cell fusions on fungal pathogenicity. For instance, for the soil-borne Fusarium oxysporum, VHF-impaired mutants exhibited only a slightly reduced virulence, whereas for the necrotrophic plant pathogen Alternaria alternata, VHFs are necessary for the full virulence of the fungus [15, 16]. Though the deletion of ZtSof1 is not essential for host penetration, colonization, or for the onset of the necrotrophic phase per se, it is during hyphal accumulation in the sub-stomatal cavity that the fusion defect impacts Z. tritici fitness. The infection process during wheat colonization was microscopically detailed before for Z. tritici [70]; however, the study did not explore the contribution of VHFs. Here, we showed that the primary intercellular hyphae surrounding the stomatal guard cells produced specialized knots from where secondary hyphae emerge and elongate to fuse with other adjacent hyphae. Hence, the preliminary hyphal network creates the basis for a symphogenous development that builds the concave-shaped pycnidial wall of the mature pycnidium. On the other hand, the inability to undergo anastomosis ceases the development of the asexual fruiting bodies at this stage, and therefore abolishes fungal reproduction. The accumulation of hyphae observed in the sub-stomata chamber by the ΔZtSof1 mutant confirms that a specific signal triggers sub-stomatal hyphal aggregation independent of the putative chemoattractant molecule secreted by the fungus to induce hyphal fusion. In line with this hypothesis, a recent study showed that the disruption of the transcription factor ZtStuA impairs hyphal aggregation, resulting in the lack of pycnidium formation in vitro and in planta [68]. Therefore, we suggest that ZtStuA acts regulating genes necessary for the pre-pycnidium stages (e.g., hyphal aggregation) [68], while ZtSof1 controls the hyphal fusions required for the formation of the mature pycnidium. The initiation of the asexual sporulation may depend on the regulatory genes, such as ZtBrlA2 and ZtFlbC, as described by Tiley et al. Moreover, the inability of ΔZtFus3 to undergo VHFs and produce pycnidia [57] confirms the interplay between the CWI and pheromone response pathways during hyphal fusion, a developmental process essential for the asexual reproduction of Z. tritici. Fleissner et al. [24] demonstrated that the deletion of So in N. crassa also affects female fertilization, blocking the sexual reproduction of the mutant. Further experiments need to be performed to address this question for Z. tritici, but, likely, the ZtSof1 gene may also play a crucial role in the sexual reproduction of this pathogen.

Conclusion

The characterization of the ZtSof1 gene demonstrated its fundamental role in fungal biology. Beyond the impact of ZtSof1 for self-fusion, we show the contribution of this gene for fungal development, including vegetative growth and melanization. Besides, we demonstrated that VHFs are dispensable for pathogenicity, but essential for pycnidial development, and these mechanisms might be controlled by the interplay between the CWI and the pheromone response pathways. Our data show how cell fusion affects Z. tritici fitness and provides a new gene target to control septoria tritici blotch (STB) disease.

Methods

Strains and growth conditions

The Swiss Z. tritici strain ST99CH_1E4 (abbreviated as 1E4), described by Zhan et al. [71], and mutant lines derived from this strain were used in this study. 1E4 strains expressing cytoplasmic GFP (1E4GFP) or mCherry (1E4mCh) were provided by Andrea Sanchez-Vallet after being generated by Sreedhar Kilaru and Gero Steinberg. The knocked-out ΔZtSlt2 [56] and ΔZtFus3 [57] mutants were provided by Marc-Henri Lebrun (National Institute of Agricultural Research – INRA, France). Because the MAPK mutants were generated in the genetic background of IPO323 [72], this strain was also used as a control. The strain IPO323 ZtHis1-ZtGFP [41] was provided by Gero Steinberg (Exeter University). Each strain was stored in glycerol at − 80 °C until required and then recovered in yeast-sucrose broth (YSB) medium (10 g/L yeast extract, 10 g/L sucrose, 50 μg/mL kanamycin sulfate; pH 6.8) incubated at 18 °C for 4 days.

Plant infection to obtain fluorescent pycnidiospores

Wheat seedlings from the susceptible wheat cultivar Drifter were grown for 16 days in the greenhouse at 18 °C (day) and 15 °C (night) with a 16-h photoperiod and 70% humidity. Blastospore suspensions of 1E4GFP or 1E4mCh were obtained after 4 days of growth in the YSB medium. Spore suspensions were adjusted to a final concentration of 106 blastospores/mL in 30 mL of sterile water supplemented with 0.1% (v/v) Tween and applied to run-off using a sprayer, and the plants were kept for 3 days in sealed plastic bags, followed by 21 days in a greenhouse. Leaves with pycnidia were harvested and transferred to a 50-mL Falcon tube containing sterile water and gently shaken to harvest the pycnidiospores. Fluorescent pycnidiospores were used to assess the cell fusion events during in vitro and in vivo growth. Pycnidiospore suspension-tagged GFP or mCherry were also adjusted to a final concentration of 106 pycnidiospores/mL, and a new batch of plants was inoculated as described above. Plants co-infected by both 1E4GFP and 1E4mCh strains were used to observe VHFs on the wheat leaf surface.

Characterization of cell fusion events in vitro and in vivo

The ability of Z. tritici to undergo cell fusions was evaluated using blastospores and pycnidiospores of 1E4GFP and 1E4mCh. Cell concentrations were adjusted to 3.3 × 107 blastospores/mL or 3.3 × 106 blastospores/mL to induce CATs or VHFs, respectively. Three hundred microliters of each morphotype and fluorescence was plated on WA to create a ratio of 1:1 and to provide a final concentration of 107 blastospores/mL or 106 blastospores/mL. A section of about 1 cm2 of agar was aseptically cut and placed on a microscope slide. The mixing of both cytoplasm contents through CATs or VHFs was checked up to 40 hai using a Leica DM2500 fluorescent microscope with LAS v.4.6.0 software. GFP excitation and emission was at 480/40 nm and 527/30 nm, respectively, whereas mCherry was excited at 580/20 nm and detected at 632/60 nm.

VHFs during spore germination on wheat leaf surfaces were obtained via confocal images using a Zeiss LSM 780 inverted laser scanning microscope with ZEN Black 2012 software. An argon laser at 500 nm was used to excite GFP fluorescence and chloroplast autofluorescence, while mCherry excitation was at 588 nm. The emission wavelength was 490–535, 624–682, and 590–610 nm for GFP, chloroplast autofluorescence, and mCherry, respectively. Plants co-inoculated with blastospores or pycnidiospores of 1E4GFP and 1E4mCh strains were checked up to 48 hai.

Plasmid constructions and transformations

Primers used for cloning, sequencing, and knock-out confirmations are listed in Table S1. DNA assemblies were conducted with the In-Fusion HD Cloning Kit (Takara BIO) following the manufacturer’s instructions. The plasmid constructions to generate the three mutants in the 1E4 genome background and used in this study (ΔZtKu70, ΔZtSof1, and ΔZtSof1-comp) are described in Additional file 13: Fig. S13. The pES1-ΔZtSof1 construction was also used to knock out the ZtSof1 gene in the 1E4GFP genome background, enabling the visualization of the GFP-tagged mutant during host infection.

Z. tritici 1E4 strain was transformed by Agrobacterium tumefaciens-mediated transformation (ATMT) according to Meile et al. [73]. The knock-out of the target genes was verified by a PCR-based approach using a forward primer specific to the upstream sequence of the disrupted gene and a reverse primer specific to bind in the resistance cassette (Additional file 14: Table S1). We determined the copy number of the transgene by quantitative PCR (qPCR) on genomic DNA extracted with the DNeasy Plant Mini Kit (Qiagen). We used qPCR target gene as the selection marker and the 18S rDNA as the reference gene (Additional file 14: Table S1). Lines with a single insertion were selected for further experiments.

Phenotypic characterizations

For all phenotypic analyses, ΔZtKu70 was considered the wild-type (WT) strain. To pinpoint the role of the ZtSof1 gene on the vegetative cell fusion, we added blastospore suspension cells of ΔZtKu70, ΔZtSof1, and ΔZtSof1-comp to a final concentration of 106 blastospores/mL or 107 blastospores/mL into WA and incubated at 18 °C to induce CATs or VHFs, respectively. For the MAPK ΔSlt2 and ΔFus3, IPO323 and ΔSlt2-complemented strains, WA plates were inoculated only at a final concentration of 106 blastospores/mL. Cell fusion events were monitored up to 40 hai by light microscopy. Because fusion bridges were not observed between individuals lacking ΔZtSof1, we mixed 150 μL of 106 blastospores/mL of ΔZtKu70, ΔZtSof1, or ΔZtSof1-comp strains with the same concentration of 1E4GFP blastospores in a ratio of 1:1 to confirm the failure of cytoplasmic streaming. At least 50 spores of each sample combination were monitored.

To test for altered fungal growth, we used PDA (39 g/L potato dextrose agar, 50 μg/mL kanamycin sulfate) or WA media to induce blastospore or hyphal growth, respectively. Two hundred microliters of spore suspension of ΔZtKu70, ΔZtSof1, and ΔZtSof1-comp was plated at a final concentration of 2 × 102 blastospores/mL on each medium as mentioned earlier and incubated in the dark at 18 °C. At least 40 colonies formed in five independent PDA plates were photographed from the bottom using a standardized camera setting [60] at 8, 9, 10, 11, 13, and 15 days post-incubation (dpi). Digital images were processed using a macro developed in the ImageJ software [74], which scores the area of individual colonies in the images. Fungal growth was obtained by converting the colony area into radial growth (mm) based on the formula \( =\sqrt{\begin{array}{c}\boldsymbol{A}\\ {}\boldsymbol{\pi} \end{array}} \). Radial growth values were plotted in a boxplot graphic using the ggplot2 package from R [75]. Analysis of variance (ANOVA) was performed to determine the differences in fungal growth among the strains using the agricolae package in R [76]. The radial growth rate (mm/day) for each strain was measured by plotting the colony radius over time, which fitted to a linear model (Pearson’s correlation coefficient value (r2 ≥ 0.98)). The relative growth rate was calculated by dividing the slope of the regression line of ΔZtSof1 by the slope of ΔZtKu70 or ΔZtSof1-comp strains. To measure mycelial growth on WA, we calculated the mycelial diameter from digital images of at least 40 colonies formed in five independent Petri dishes at 15 dpi and using ImageJ software [74]. The mycelial diameter values were divided by two to generate the radial growth (mm) values.

The effect on the blastosporulation was assessed by adding each tested Z. tritici strain at an initial concentration of 104 blastospores/mL into YSB and incubated at 18 °C. Aliquots of each flask were taken every 24 h for 4 days, and the blastospore concentration was calculated using a KOVA cell chamber system (KOVA International Inc., USA).

The degree of melanization was estimated from at least 40 colonies formed on PDA plates. We used a macro developed in ImageJ [74], which scores the mean gray value of each colony. Gray values range from 0 to 255, with 0 representing black and 255 representing white. The mean gray values of each strain over time were plotted in a boxplot using ggplot2 package from R [75].

The impact on the cell integrity caused by the deletion of ΔZtSof1 was verified by exposing the blastospores of ΔZtKu70, ΔZtSof1, and ΔZtSof1-comp to nine different stress conditions, including different temperatures (18 °C and 27 °C), oxidative stress (0.5 and 1 mM of hydrogen peroxide (H2O2)), osmotic stress (1 M sodium chloride (NaCl) and 1 M sorbitol), cell wall stress (2 mg/mL Congo red and 10 μg/mL Calcofluor white (CFW)), and plasma membrane stress (0.01% sodium dodecyl sulfate (SDS)). Spore suspensions of each strain were serial diluted to 4 × 104, 4 × 105, 4 × 106, and 4 × 107 blastospores/mL, and drops of 3.5 μL were plated on five independent PDA plates amended with the mentioned stresses and incubated at 18 °C. Colony phenotypes were assessed by digital images taken at 5 dpi.

Virulence assay and pycnidium formation in vitro

Seeding, greenhouse and plant growth conditions, inoculum preparation, and plant inoculation followed the procedures described by Meile et al. [73]. To estimate the percentage of leaf covered by lesions (PLACL) and pycnidium formation, we harvested the second leaves of Drifter plants inoculated with ΔZtKu70, ΔZtSof1, or ΔZtSof1-comp strains at 8, 10, 11, 12, 14, 16, and 21 dpi. Leaves were mounted on a paper sheet, scanned with a flatbed scanner, and analyzed using automated image analysis [77]. Data analysis and plotting were performed using the ggplot2 package [75].

The defect in pycnidium formation was confirmed by plating blastospores of ΔZtKu70, ΔZtSof1, and ΔZtSof1-comp strains onto wheat extract agar medium (50 g/L blended 21-day-old wheat leaves cultivar Drifter, 10 g/L agar) and incubated under UV-A light (16:8 light:dark cycle) up to 40 days at 18 °C, following the protocol optimized by Tiley et al. [68].

Confocal laser scanning microscopy of infected wheat leaves

To assess the impact of ZtSof1 deletion on fungal fitness during host colonization, we inoculated wheat plants with 1E4GFP and other two independent GFP-tagged ΔZtSof1 mutants. Infected leaves were harvested at 6, 7, 8, 9, 10, 11, and 12 dpi and checked for developmental stages of asexual fruiting bodies. Microscopy was conducted using Zeiss LSM 780 inverted laser scanning microscope with ZEN Black 2012 software. An argon laser at 500 nm was used to excite GFP fluorescence and chloroplast autofluorescence with an emission wavelength of 490–535 nm and 624–682 nm, respectively. Analyses, visualization, and processing of image z-stacks were performed using ImageJ software [74].