Summary
Although lipid signaling has been shown to serve crucial roles in mammals and plants, little is known about this process in filamentous fungi. Here we analyse the contribution of phospholipase D (PLD) and its product phosphatidic acid (PA) in hyphal morphogenesis and growth of Epichloë festucae and Neurospora crassa, and in the establishment of a symbiotic interaction between E. festucae and Lolium perenne. Growth of E. festucae and N. crassa PLD deletion strains in axenic culture, and for E. festucae in association with L. perenne, were analysed by light-, confocal- and electron microscopy. Changes in PA distribution were analysed in E. festucae using a PA biosensor and the impact of these changes on endocytic recycling and superoxide production investigated. We found that E. festucae PldB and the N. crassa ortholog, PLA-7, are required for polarized growth, cell fusion and ascospore development, whereas PldA/PLA-8 are dispensable for these functions. Exogenous addition of PA rescues the cell-fusion phenotype in E. festucae. PldB is also crucial for E. festucae to establish a symbiotic association with L. perenne. This study identifies a new component of the cell-cell communication and cell fusion signaling network that controls hyphal morphogenesis and growth in filamentous fungi.
Introduction
Epichloë festucae is a filamentous ascomycete fungus that forms a largely beneficial interaction with the agriculturally important grass Lolium perenne (Schardl, 2010, Tanaka et al., 2012). This endophyte colonises the apoplast of aerial tissues of the infected plant to form a restricted hyphal network in planta (Christensen et al., 2008, Scott et al., 2012). E. festucae also forms an epiphyllous hyphal network on the leaf surface that is connected to the endophytic network following exit of these hyphae through the cuticle by formation of an appressorium-like structure called an expressorium (Becker et al., 2016).
Growth of the fungus in planta is highly regulated and many well-known signaling molecules and pathways are required for a functional interaction. These include signaling via reactive oxygen species (ROS), the cell wall integrity (CWI) mitogen-activated protein kinase pathway (MAPK) and the stress-activated MAPK pathway (Becker et al., 2015, Eaton et al., 2010, Tanaka et al., 2006). Many signaling pathways require interplay between membrane-bound and cytosolic proteins. In mammals and plants, this frequently involves lipid signaling via different lipid species which target various proteins to the membrane. While studies on lipid signaling have received considerable attention in mammalian cells, plants and yeasts, few studies have focused on the role of this process in filamentous fungi. Phosphatidic acid (PA), produced by phospholipase D (PLD)-catalysed hydrolysis of phosphatidyl choline (PC), is an important second messenger in plant and mammalian cells, where it is required for the correct temporal and spatial localization and activity of many proteins (Selvy et al., 2011, Jenkins & Frohman, 2005). Through these interactions PA influences cell growth, cytoskeletal rearrangement, vesicle formation and trafficking as well as superoxide production (Selvy et al., 2011, Wang, 2005). In plants, PA is additionally involved in pathogen and stress responses, stomatal opening and in signaling by the hormone, abscisic acid (Zhang et al., 2009, Yao & Xue, 2018). PA also has an important role in both endo- and exocytosis vesicle trafficking. Specifically, PA is involved in vesicle transport from the endoplasmic reticulum (ER) to the Golgi apparatus, including vesicle budding from the Golgi apparatus via ADP ribosylation factor Arf1 signaling and epidermal-growth factor (EFR)-mediated endocytosis (Chen et al., 1997, Ktistakis et al., 1996, Shen et al., 2001). Furthermore, in chromaffin cells PA accumulates at sites of exocytosis where it is required for the formation of fusion-competent granules (Tanguy et al., 2019, Zeniou-Meyer et al., 2007). In mammalian and plant cells, PA also contributes to superoxide production by activating the multi-protein NADPH oxidase (Nox) complex, via both the direct activation of the catalytic subunit as well as through the recruitment of the regulatory subunits. Depletion of PLDα in Arabidopsis leaves and the inhibition of PLD in tobacco pollen tubes reduces NADPH oxidase activity (Potocký et al., 2012, Sang et al., 2001). In addition, the direct interaction between RBOHD and PA is required to achieve stomatal closure upon abscisic acid signaling (de Jong et al., 2004, Zhang et al., 2009). In animal cells PA has been found to interact with multiple components of the Nox complex, such as the catalytic subunit gp91phox (or Nox2) (Kanai et al., 2001, Karathanassis et al., 2002, Ago et al., 2003, Qualliotine-Mann et al., 1993, Taylor et al., 2004, Taylor et al., 2012). In a cell-free system, PA was shown to be crucial for the activation of the mammalian NADPH oxidase (Qualliotine-Mann et al., 1993, Palicz et al., 2001).
The role of PA in yeast and filamentous fungi has been analysed by studying the phenotype of PLD deletion strains. In Saccharomyces cerevisiae, deletion of SPO14, the homolog of mammalian PLD1, results in a loss of fusion of Golgi-derived vesicles to the prospore membrane during sporulation (Rudge et al., 1998, Riedel et al., 2005, Nakanishi et al., 2006). In addition, Spo14 is necessary for morphogenesis and polarized growth in response to pheromone and Sec14-independent secretion (Hairfield et al., 2001, Harkins et al., 2008, Rudge et al., 2002). In Candida albicans, PLD1 is necessary for the dimorphic transition and full virulence (McLain & Dolan, 1997, Dolan et al., 2004). In filamentous fungi, the role of PLDs to date has only been analysed in Aspergillus fumigatus and Fusarium graminearum. A. fumigatus has three PLD isoforms, one of which is required for internalization of the fungus into host epithelial cells and for full virulence in certain immunosuppressed mice (Li et al., 2012). F. graminearum, also has three pld genes, but only one, Fgpld1, is crucial for vegetative development and host virulence. Deletion of Fgpld1 resulted in reduced colony size, defects in sporulation and sexual development, decreased production of the mycotoxin deoxynivalenol (DON), and reduced virulence on wheat (Ding et al., 2017). The molecular mechanisms behind these phenotypic changes however, remain unclear.
The genome of E. festucae encodes two putative PC-hydrolysing PLD genes, which we have named pldA and pldB. The aim of this study was to examine the role of these two proteins, and their orthologs from the model organism Neurospora crassa, in hyphal morphogenesis, growth and development in axenic culture, and to test if they are required for establishment of a mutualistic symbiotic association between E. festucae and L. perenne.
Results
The E. festucae genome encodes two putative phosphatidylcholine-hydrolysing phospholipase D enzymes
To identify PLD-encoding genes in the E. festucae Fl1 genome, a tBLASTn analysis was performed using S. cerevisiae Spo14 as a query sequence. This search identified two putative homologs: EfM3.055250 (E-value of 1e-89) and EfM3.032570 (E-value of 0), hereafter referred to as pldA and pldB, respectively. An analysis of previous transcriptome datasets showed that both genes are moderately expressed in axenic culture (pldA: 10.39 RPMK, pldB: 35.96 RPMK) (Hassing et al., 2019). The expression of pldB is downregulated two-fold in planta relative to axenic culture, while the expression of pldA is not significantly different (Hassing et al., 2019). PldA and PldB are 877 and 1,713 amino acid residues in length, respectively (Fig. 1A) and share 39.8% amino acid sequence identity. Both sequences contain four conserved domains characteristic of PLDs, including the HxKx4D(x6GSxN) (HKD) motif in domains II and IV (Ponting & Kerr, 1996, Koonin, 1996), necessary for phosphatidyltransferase activity (Sung et al., 1997, Liscovitch et al., 1999, Liscovitch et al., 2000). PldB in addition contains a sequence homologous to a PtdIns[4,5]P2-binding polybasic motif located between domains II and III (Sciorra et al., 1999, Sciorra et al., 2002) (Fig. 1A). Analysis with InterproScan (Jones et al., 2014) and SMART (Letunic et al., 2015) identified slightly overlapping, N-terminal phosphoinositide-binding PX and PH-domains in PldB but not in PldA. Both domains are present in mammalian and fungal PLDs, but only in the Pldζ group of PLDs found in plants (Selvy et al., 2011).
Phylogenetic analysis of all PLD sequences available at Ensembl (Zerbino et al., 2018), revealed a separation between plant and animal/fungal PLDs, presumably due to the presence of the different protein domains. Using only selected sequences we observed a separation between PldB-like and PldA-like PLDs within the fungal kingdom, demonstrating that both isoforms are conserved among filamentous fungi. All analysed fungal genomes with the exception of E. festucae, contained a third PLD isoform, which separated as a subgroup of PldA-like PLDs, here named PldA2-like (Figs. 1B & 1C).
An amino acid sequence alignment of PldA and PldB with the corresponding homologs from N. crassa, M. oryzae, F. graminearum and A. fumigatus, revealed that PldA and PldB homologs have high sequence identity (an average of 68% and 58%, respectively) (Fig. S1A). For PldB, the sequences were most similar within the catalytic core (Fig. S1B).
PldB is required for hyphal morphogenesis, growth and fusion
To investigate whether E. festucae PldA and PldB are required for hyphal morphogenesis and growth, each gene was individually deleted from the WT background by gene replacement. Three hygromycin-resistant ΔpldA strains (#T47, #T49, #T82) and three geneticin-resistant ΔpldB strains (#T26, #T50, #T87) were identified (Fig. S2).
Although deletion of pldA had no obvious effect on colony growth and morphology, deletion of pldB led to a significant reduction in colony size and an increase of aerial hyphae (Fig. 2). Analysis by differential interference contrast (DIC) light microscopy confirmed that the hyphal morphology of ΔpldA strains was indistinguishable from WT (Fig. 2B-F). In contrast, ΔpldB strains frequently had swollen and highly vacuolated hyphae and hyphal tips (Fig. 2F). Staining of the colony with the chitin stain CFW revealed occasional accumulation of the stain in the cytoplasm of hyphae, which is indicative of a defect in the recycling of cell wall components (Higuchi et al., 2009). Intra-hyphal hyphae (IHH) were also occasionally detected. Although both strains formed hyphal bundles and coils like WT, no cell fusion could be observed in the ΔpldB strains (Figs. 2E & S3). To verify this cell fusion defect, ΔpldB strains expressing either GFP or mCherry under the regulation of a constitutive promoter, were grown in close proximity to one another and analysed for cytoplasmic mixing (Becker et al., 2015). WT hyphae with both GFP and mCherry fluorescence were frequently observed, but in the ΔpldB strains the hyphae had either GFP or mCherry fluorescence but not both (Fig. 3). Addition of PA to the growth medium rescued the ΔpldB hyphal cell fusion defect (Figs. 4 & S4), although this chemical complementation with PA was concentration dependent. However, PA was unable to rescue the hyphal morphology and growth defect of ΔpldB. Introduction of the WT pldB allele into ΔpldB did complement all observed axenic culture phenotypes (Figs. 2 & S3) confirming that the defects were due to the specific deletion of the pldB gene.
To determine whether there was any functional redundancy between PldA and PldB, double deletion strains were generated by deleting pldA in the ΔpldB#T26 and #T87 backgrounds. Two strains with a pldA deletion were obtained (ΔpldBΔpldA#T3 and ΔpldBΔpldA#T5) (Fig. S2) and both had axenic culture phenotypes comparable to the single ΔpldB strains (Fig. 2).
To determine whether overexpression of these genes would influence hyphal morphology and growth, constructs with pldA and pldB under the control of the A. nidulans gpdA promoter were transformed into the WT strain. Strains with a range of overexpression values (pldA OE strains: #T3, #T4, #T9, pldB OE strains: #T6, #T21, #T22; Fig. S5), as determined by RT-qPCR, were grown alongside WT in axenic culture but found to have no obvious phenotype differences (Fig. 2).
The PldB ortholog PLA-7 is required for hyphal fusion and spore production in N. crassa
To test if the role of PldB in vegetative cell fusion is conserved in other fungi, mutants carrying a deletion of the genes homologous to PldA (NCU10400, PLA-8) and PldB (NCU03955, PLA-7) were obtained from the N. crassa gene knock-out collection. While strains are usually available as homokaryotic isolates, the Δpla-7 mutant was recorded in the collection as heterokaryotic. This indicated that pla-7 is either essential or required for successful sexual crossing. Single spore isolation resulted in homokaryotic strains, indicating that pla-7 is not essential.
In N. crassa vegetative cell fusion occurs at two stages of mycelial development: at an early stage when germinating conidia (germlings) home towards each other and undergo consecutive fusion events to form supracellular structures, and at a later stage, within mature colonies, when hyphal branches fuse to form cross connections that increase the interconnectivity of the mycelium. The cell fusion phenotype of germinating conidia or mature hyphae of the Δpla-8 mutant was indistinguishable from the WT strain. In contrast, germination of Δpla-7 mutant spores, was significantly delayed, and germling homing and fusion was completely absent. Addition of PA, using the same concentration range as used for E. festucae, failed to rescue the germling fusion phenotype of N. crassa Δpla-7 (Fig. S6). No hyphal fusion was detected in mature Δpla-7 mutant colonies (Fig. 5). Together these observations indicate that PLA-7 has a conserved essential function in vegetative cell fusion.
Unlike E. festucae, N. crassa readily forms sexual structures in axenic culture thereby enabling the role of PLA-7 in sexual development to be analysed by crossing Δpla-7 with the WT strain. While the Δpla-7 strain failed to produce female reproductive structures, crosses between WT as a female and Δpla-7 as the male resulted in ascospores. However, ascus formation was aberrant and sometimes asci contained less than the usual eight ascospores. Overall WT/Δpla-7 crosses produced significantly fewer spores than WT/WT pairings (Fig. 5). In addition, the fraction of mutant progeny was just 17%, compared to an expected 50% for a 1:1 segregation of Δpla-7 to pla-7+in this cross.
PldB localises to septa and the plasma membrane
PLDs analysed to date localise to various cellular compartments, including the Golgi apparatus, vesicles, endosomes, the plasma membrane (PM), and the cytosol (Colley et al., 1997, Hughes & Parker, 2001, Du et al., 2003, Selvy et al., 2011). To investigate the localization of E. festucae PldA and PldB, N-terminal GFP translational fusion constructs were generated and transformed into WT protoplasts. Expression of fusion proteins of the expected size was verified by western blot analysis (Fig. S7). To confirm the functionality of the PldB localization constructs, the constructs were transformed into ΔpldB strains, which restored WT-like growth in axenic culture (Fig. S8).
The PldA-GFP fusion protein localized to the cytoplasm in hyphae of different developmental stages, a result comparable to that of strains expressing cytosolic GFP driven by the same promoter (Fig. 6). The PldB-GFP fusion protein was also found to localise to the cytoplasm of hyphae of all ages. However, in young hyphal bundles, as well as in mature hyphae, the signal was also detected in septa and occasionally associated with the PM (Fig. 6), a result indicative of a transient localization of PldB to the membrane (Selvy et al., 2011).
PldB is crucial for the E. festucae-L. perenne symbiotic interaction and expressoria formation
To analyse the role of PldA and PldB in the interaction of E. festucae with its host, deletion and overexpression (OE) strains were inoculated into seedlings of L. perenne and the phenotype of the mature plants examined at 10 weeks post-infection. Plants infected with ΔpldA, or pldA and pldB OE strains, were indistinguishable from those infected with WT. In contrast, plants infected with ΔpldB strains were severely stunted, with tiller lengths significantly less than that of plants infected with WT (Fig. 7). While the tiller number of plants infected with ΔpldB#T87 was significantly greater than WT, there were no significant differences in tiller number for the other two, independent ΔpldB mutants. Introduction of a pldB WT allele into the ΔpldB strains restored the WT plant interaction phenotype. For the ΔpldBΔpldA double deletion strain tiller number was increased and tiller length reduced, compared to WT (Fig. 7).
To examine hyphal growth in planta, pseudostem cross sections were fixed and embedded in resin, stained with toluidine blue, and the number of hyphae per intercellular space determined by light microscopy. Plants infected with the ΔpldB strains had a significantly greater number (P < 1E-10) of hyphae per intercellular space (3.02) than plants infected with WT strain (1.28). The number of hyphae per intercellular space for plants infected with ΔpldA (1.27), pldA OE (1.37) and pldB OE (1.27) strains were not significantly different to WT (Figs 8A & S9). TEM ultrastructure analysis of pseudostem cross sections revealed that the ΔpldB and ΔpldBΔpldA double deletion strains occasionally colonise the vascular bundles, a phenotype seldom observed for WT (Fig. 8B-E). Hyphae from either of these two mutants were often less electron-dense or completely vacuolated and frequently formed IHH when growing in the intercellular space (Fig. 8B-E). Another distinct phenotype of both mutants was their apparent reduced ability to degrade the plant cuticle compared to WT and the ΔpldA strains, leading to a proliferation of hyphae immediately below the cuticle surface (Fig. 8).
To further investigate the cellular phenotype of the host interaction, longitudinal sections of the pseudostem were infiltrated with WGA-AF488, which binds chitin, and aniline blue, which binds β-glucans, and analysed by confocal laser scanning microscopy. The ΔpldA, and pldA and pldB OE strains had a hyphal morphology and restricted growth phenotype in planta indistinguishable from WT (Fig. 9). In contrast, growth of the ΔpldB strains was unrestricted with multiple hyphae in the intercellular space, which formed bundles and multicellular lobed structures. Furthermore, ΔpldB strains were unable to efficiently form expressoria, resulting in a proliferation of hyphae immediately below the cuticle. However, these mutant hyphae were still able to occasionally breach the cuticle to grow as epiphytic hyphae on the surface of the leaf (Fig. 9). The ΔpldBΔpldA double mutant had a very similar hyphal morphology and growth phenotype in planta to the ΔpldB strain. As expected, the in planta hyphal morphology and growth defects were rescued by introduction of a WT allele of pldB. (Fig. 9).
Deletion of pldB, but not pldA, results in delocalization of a phosphatidic acid biosensor from the plasma membrane
To analyse the effect of the pldA and pldB deletions on the localization of PA in the cell, a PA biosensor was constructed comprising a Spo20 PA-binding domain fused to GFP (Potocký et al., 2014). This construct (pBH50) was transformed into WT, ΔpldA and ΔpldB strains, and the resulting transformants shown to express fusion proteins of the correct molecular weight (Fig. S10). In the WT strain, the PA biosensor localised to the PM and septa in young as well as mature hyphae (Fig. 10). Interestingly, the signal in newly formed hyphal branch tips was more intense than in the tips of hyphae at the colony growth front. PA localization in the ΔpldA strains was indistinguishable from WT. However, in the ΔpldB strains, the signal was largely cytoplasmic, and localization to the PM and septa was only detected in mature hyphae (Fig. 10A). In addition, the PA biosensor accumulated in an organelle in the cytoplasm of the ΔpldB strains that was shown by DAPI staining to be the nucleus (Fig. 10B). To examine the localization of PA in fungal hyphae in planta, a WT strain expressing the PA biosensor was inoculated into L. perenne seedlings and GFP expression examined in mature plants. GFP fluorescence was detected evenly along the PM and septa of both endophytic as well as epiphytic hyphal tips and mature hyphae, a pattern identical to that observed for PA localization in hyphae in axenic culture (Fig. 10C).
The absence of PldB has no effect on localization of Nox complex components
In mammalian systems PA produced by PLD has been shown to activate the Nox complex by promoting the recruitment of the Nox complex subunits to the membrane as well as by direct interaction of PLD1 or PLD2 with Nox-activating proteins (Regier et al., 2000, Kanai et al., 2001, Karathanassis et al., 2002, Ago et al., 2003, Chae et al., 2008, Selvy et al., 2011, Jang et al., 2012). We therefore tested for a direct interaction between PldB and components of the Nox complex (NoxR, BemA, RacA, Cdc24) and other potential regulators (Cdc42, PkcA (EfM3.006500), MssD (EfM3.031950)) by yeast two-hybrid assays. However, PldB did not interact with any of these proteins (Fig. S11). We next examined whether deletion of pldA or pldB would affect the localization of components of the Nox complex, including the catalytic transmembrane protein NoxA, the regulatory proteins NoxR, BemA, and Cdc24, and the small GTPase RacA using GFP translational fusions of each (Fig. S12) (Takemoto et al., 2011). In the WT background all fusion proteins localised as previously reported and this pattern of localization was unchanged in both the ΔpldA and ΔpldB strains (Fig. S12).
Absence of PldB results in mislocalized reactive oxygen species production
While the localization of the components of the NoxA complex was not altered, PA produced by PLD has been shown to directly activate Nox complexes in mammalian and plant systems (Karathanassis et al., 2002, Taylor et al., 2004, Taylor et al., 2012, Potocký et al., 2012). To investigate the role of PA in the activation of fungal NADPH oxidases, the production and localization of superoxide was analysed in ΔpldA and ΔpldB strains in culture, using NBT and a ROS-ID® Superoxide kit.
When WT colonies were stained with NBT, a dark blue formazan precipitate was formed that was predominantly localised to hyphal tips, with only a small number (∼5%) having either an aberrant staining pattern or no detectable signal (Fig. 11). A similar pattern of staining was observed for the ΔpldA strains, as well as the pldA and pldB OE strains. In comparison, this aberrant pattern of NBT staining of tips was increased to approx. 20% in ΔnoxA, and approx. 40% in ΔpldB. Normal tip staining was restored in the ΔpldB/pldB complementation strains (Fig. 11).
Hyphae were also stained with the ROS-ID® superoxide probe. In WT tips the probe localised to puncta and in mature hyphae the dye accumulated in long filaments containing more intense puncta. Given these structures colocalised with the mitochondrial Reagent-Green marker, the superoxide detected appears to be generated as a by-product within mitochondria. Similar observations were made with ΔpldA strains. For the ΔpldB strains the ROS probe localised to long filamentous structures in tips, which, with the exception of mature hyphae, were not seen in WT or the ΔpldA mutant. These ROS positive structures also colocalised with the mitochondrial Reagent-Green marker demonstrating they are also mitochondria. It is likely that these unique distributions reflect distorted mitochondria and result from excessive vacuole formation characteristic of the ΔpldB strains (Fig. 11).
The deletion of pldB may influence endocytic recycling
As mammalian PLD1 and PLD2 as well as PA have roles in both endo- and exocytosis, the vesicle trafficking phenotypes of the ΔpldA and ΔpldB mutants were examined by introducing GFP fusions of a suite of proteins known to localize to specific structures/stages of secretion and endosomal recycling, including: late Golgi cisternae (Vps52), the Spitzenköerper (Rab11), the exocyst (Sec3, Exo70), the endocytic collar (fimbrin) and early endosomes (Rab5) (Kilaru et al., 2015, Guo et al., 2015, Sánchez-León et al., 2015, Upadhyay & Shaw, 2008).
In WT strains, Vps52 localized to small, mobile puncta. While the apical region was mainly free of these, small accumulations or aggregates at the very apex of the tip were occasionally observed. Rab11 assumed a far more discreet distribution and was observed exclusively as a spot at the hyphal tip, presumably marking the Spitzenköerper. Although cytoplasmic signals could make clear identification difficult in some instances, the exocyst component Exo70 also collected at the hyphal tip but in a broader crescent shape than Rab11. Sec3 was observed as a crescent that emanated from a spot-like concentration also at the hyphal tip. These distributions differed markedly from the endocytic collar protein Fimbrin, which appeared as puncta that associated at the membrane and to a much lesser extent the cytoplasm, proximal to, but not at, the hyphal apex. An additional endocytic protein, Rab5, was also absent from the tip. This protein appeared as different sized accumulations which were motile (Fig. 12). The localization of Vps52, Rab11, Exo70, Sec3 and Rab5 GFP fusion proteins in ΔpldA and ΔpldB strains was indistinguishable from WT (Fig. 12). However, the characteristic sub-apical collar localization of fimbrin in WT was frequently absent in the ΔpldB mutant with the protein instead concentrating along the apical tip as a robust and broad crescent (Figs. 12 & S13).
Discussion
Membrane-bound conversion of phosphatidyl choline (PC) to phosphatidic acid (PA) has an important role for morphogenesis, growth and development in animals, plants and yeast, but very little is known about the importance of this process in filamentous fungi (Selvy et al., 2011). Here we show that phospholipase D (PLD)-dependent synthesis of PA is required for cell fusion in both E. festucae and N. crassa, highlighting the importance of lipid signaling for the fusion process. Cell fusion in fungi is crucial for hyphal network formation and multi-cellular development enabling nutrient transfer and signaling. This lipid conversion is also vital for development of a mutualistic symbiotic association between E. festucae and L. perenne.
The genome of E. festucae encodes two putative PC-hydrolysing PLDs, PldA and PldB, which differ significantly in their domain structure. PldB has PH and PX domains, which are also present in yeast, mammalian and plant Pldς homologs (Selvy et al., 2011). These domains bind specific phosphoinositides that are required for localization of the protein (Selvy et al., 2011, Stahelin et al., 2004, Sciorra et al., 2002, Hodgkin et al., 2000). PldB also contains a PtdIns[4,5]P2-binding polybasic motif domain that is present in many PLDs of higher eukaryotes (Selvy et al., 2011). This motif is required for activation of the protein, but not for its localization (Sciorra et al., 1999, Sciorra et al., 2002). All of these domains are absent in PldA, although this isoform has retained the phosphatidyltransferase activity domains. Consistent with this domain structure, PldB localises to lipid-rich structures, in septa and the PM, while PldA localized exclusively to the cytosol of the hyphae. S. cerevisiae Spo14 localizes to the cytoplasm, to endosomes, the nucleus, and is recruited to the prospore membrane during sporulation (Rudge et al., 1998, Li et al., 2000, Garcia-Lopez et al., 2011). Recruitment of mammalian PLD1 and PLD2 to perinuclear membranes, the Golgi apparatus, endosomes, the PM, and membrane ruffles was found to be highly dependent on the activation status and developmental stage (Colley et al., 1997, Du et al., 2003, Hughes & Parker, 2001, Selvy et al., 2011). Thus, PldA and PldB may also assume other distributions during different stages of development.
Deletion of pldB resulted in reduced colony growth and increased aerial hyphae, phenotypes similar to those observed in F. graminearum Δpld1 strains (Ding et al., 2017). Microscopic analysis of the ΔpldB strains in culture showed that hyphae were often variable in diameter, abnormally shaped, frequently swollen and had an uncommonly high degree of vacuolation. Intra-hyphal hyphae were also occasionally detected. In contrast, growth of ΔpldA, pldA OE and pldB OE were indistinguishable from WT as found for mutants of the pldA homologs in A. nidulans (AnpldA) and F. graminearum (Fgpld2) (Ding et al., 2017, Hong et al., 2003). These observations highlight the importance of PldB, but not PldA, for normal polarized growth. While genetic complementation was able to restore the WT-like phenotype, the exogenous addition of PA did not, indicating that PldB itself is required for normal hyphal growth. Interestingly, similar phenotypes were observed for the E. festucae ΔracA mutant (Tanaka et al., 2008, Kayano et al., 2018). RacA is a small GTPase, which contributes to actin polymerization and Nox complex activation and has been found to interact with PLDs in mammalian cells (Jang et al., 2012, Hess et al., 1997, Bishop & Hall, 2000). Indeed, PLD2 has been found to act as a guanine exchange factor (GEF) for mammalian Rac2 (Mahankali et al., 2011).
Deletion of either pldB in E. festucae or pla-7 in N. crassa abolished hyphal fusion in axenic culture, demonstrating that PldB/PLA-7 is a newly discovered component of the cell-to-cell communication and cell fusion signaling network between cells of Ascomycete filamentous fungi (Read et al., 2010, Fischer & Glass, 2019). To date, approximately 70 mutants with a fusion defect have been identified, and among these are proteins required for the cell wall integrity and pheromone response MAP kinase pathways, the STRIPAK complex, ROS signaling, as well as phosphatases and transcription factors (Fischer & Glass, 2019). The first step in the cell fusion process is chemotropic growth of germlings or hyphae towards each other, followed by degradation of the cell wall and fusion of membranes. Because the PA molecule has a cone shape, it can influence membrane curvature and could therefore be mechanistically required for membrane fusion during hyphal cell fusion (Selvy et al., 2011, Fischer & Glass, 2019). However, in Δpla-7 mutants, chemotropic growth was not observed, indicating that instead of a requirement of PA for membrane fusion, PA may actually be required for signaling during chemotropic growth. Interestingly, defects in ergosterol biosynthesis also result in cell-to-cell communication and cell fusion defects (Weichert et al., 2016). Nevertheless, exogenous addition of PA was able to restore hyphal fusion to the E. festucae ΔpldB strains. Whether this chemical complementation can be attributed to the restoration of signaling pathways required for chemotropic growth or changes in the ability of the membranes to fuse, remains to be determined. In mammals, PA interacts with, and recruits, proteins such as Rho and Arf GTPases, which may influence vesicle transport and actin cytoskeleton organization (Jang et al., 2012). Indeed, formation of an actin cytoskeleton is an important component of the hyphal fusion process as shown by treatment of N. crassa with the actin-polymerization inhibitor latrunculin B, which abolished hyphal fusion (Roca et al., 2010). The PLD proteins also bind actin and interact with a multitude of proteins in mammals (Lee et al., 2001, Kusner et al., 2002), including the Rho family GTPases, CDC42 and RAC1, the homologs of which are required for cell polarity and hyphal fusion in N. crassa, respectively (Lichius & Lord, 2014, Araujo-Palomares et al., 2011). In S. cerevisiae Cdc42 activates Spo14 (PLD1) and PA activates the p21-activated kinase Ste20 during polarized growth in response to pheromone (Harkins et al., 2008). While our yeast two-hybrid analysis did not reveal interactions between E. festucae PldB and Rac1 or Cdc42, these results do not exclude the possibility as detecting interactors with the yeast homolog, Spo14 (PLD1), is inherently difficult (Riedel et al., 2005). Whether the lack of fusion observed here is due to a defect in the ability of PLD or PA to interact with the actin cytoskeleton or on other signaling pathways required for fusion such as the CWI MAP kinase pathway or ROS signaling, remains to be elucidated.
Deletion of the pldB homolog in N. crassa resulted in a female-sterile phenotype, characterised by the failure to produce pre-fruiting bodies. Only crosses where the female is WT resulted in the formation of mature perithecia and ascospores, a phenotype commonly observed for other fusion-negative mutants (Fischer and Glass, 2019). This developmental stage could not be analysed in E. festucae as it only undergoes sexual reproduction in association with the host plant (Bultman & Leuchtmann, 2009). A defect in perithecia formation was also observed for the F. graminearum Δpld1 strain (Ding et al., 2017).
In N. crassa, asci resulting from a cross between a Δpla-7 male and a WT female, were reduced in number and malformed, a phenotype not commonly observed for fusion mutants. N. crassa possesses a mechanism called meiotic silencing by unpaired DNA (MSUD), in which the RNA generated from unpaired DNA at meiosis is silenced (Shiu et al., 2001, Hammond, 2017). As pla-7 is absent in the genome of the male partner in these crosses, MSUD likely results in the complete silencing of pla7 expression during meiosis. Therefore, these results suggest that PLA-7 has a role in ascus and/or ascospore development, as has been show for Spo14, the homolog from S. cerevisiae (Rudge et al., 1998). While meiosis itself is not affected, deletion of SPO14 results in the inability of vesicles to fuse to the prospore membrane due to the loss of activation and recruitment of the PA-binding t-SNARE Spo20 to the prospore membrane (Connolly & Engebrecht, 2006, Nakanishi et al., 2006). However, while Spo14 is absolutely essential for spore formation, PLA-7 is not, as some viable ascospores were obtained in these crosses, although at a much lower frequency of segregation (17%) than expected (50%). Whether this outcome correlates with the silencing of Pla7 during meiosis or is due to other functions of PLA-7 or PA remains to be determined.
PLA-7 also appears to have a role in the asexual life cycle of N. crassa, as conidiospore germination is delayed in the Δpla-7 mutant, a phenotype also observed for the F. graminearum Δpld1 mutant (Ding et al., 2017). As E. festucae strain Fl1 conidiates very poorly it was difficult to assess this phenotype in this species (Becker et al., 2015). These results highlight the importance of PLA-7 for both the asexual and sexual lifecycles of N. crassa.
The severe host interaction phenotype observed for the ΔpldB mutant was very similar to that reported previously for several other E. festucae symbiotic mutants (Tanaka et al., 2006, Tanaka et al., 2013, Eaton et al., 2010, Becker et al., 2015, Takemoto et al., 2006, Green et al., 2017). Instead of wild-type like restricted hyphal growth, the ΔpldB mutant had a proliferative pattern of growth in planta. There was an increase in the number of hyphae in the host intercellular spaces of the leaves and colonization of the host vascular bundles, phenotypes not seen in WT associations. A transcriptome analysis of E. festucae symbiosis-defective mutant strains in association with L. perenne revealed an upregulation of genes in primary metabolism, host cell wall degradation, and peptide and sugar transporters; changes indicative of a starvation response (Eaton et al., 2015, Chujo et al., 2019). Indeed, all fusion-negative E. festucae strains obtained to date have a stunted host phenotype (Scott et al., 2018). However, given E. festucae mutants in cAMP/PKA signaling (Voisey et al., 2016), apoplastic iron homestasis (Johnson et al., 2013), and chromatin regulation (Chujo et al., 2019), exhibit a proliferative growth phenotype in planta but are not defective in cell fusion, an alternative more general hypothesis is needed to explain these results. One possibility is alkalinisation of the apoplast (Scott et al., 2018); a change that activates signaling through the PacC-mediated pH responsive and MAP kinase pathways, triggering invasive growth (Fernandes et al., 2017).
In addition to the hyphal fusion defect, the ΔpldB mutant was unable to form expressoria, appressorium-like structures that allow endophytic hyphae below the cuticle to swell up and penetrate this layer to emerge and grow on the surface of the leaf. Instead, the hyphae proliferate below the surface of the cuticle, a phenotype also observed for E. festucae nox complex, and some other symbiotic mutants that are defective in a conserved signaling network for cell fusion and multi-cellular development (Becker et al., 2016, Green et al., 2017, Green et al., 2019, Scott et al., 2018). Development of a functional appressorium in M. oryzae requires both Nox1 and Nox2 (Egan et al., 2007). Nox2 is necessary for septin-mediated actin ring reorientation at the septal pore, and Nox1 for maintenance of the F-actin network during penetration and elongation (Ryder et al., 2013). Given the similarity in expressorium phenotypes of the E. festucae ΔpldB and nox complex mutant strains, PldB or the PA product may be involved in Nox complex assembly and/or activation leading to a change in the actin cytoskeleton network. Indeed, PA has been shown to be involved in the recruitment of the cytosolic components of the Nox complex in plant and mammalian systems (Regier et al., 2000, Kanai et al., 2001, Karathanassis et al., 2002, Ago et al., 2003, Chae et al., 2008). However, deletion of either pldA or pldB in E. festucae did not alter the cellular localization of the Nox complex components, NoxA, NoxR, BemA, Cdc24 or RacA.
PA has also been shown to stimulate mammalian NADPH oxidase activity in a cell-free system through direct interaction with gp91phox, and in plant cells, by direct interaction with the NADPH oxidase (Palicz et al., 2001, Qualliotine-Mann et al., 1993, Taylor et al., 2004, Taylor et al., 2012, Zhang et al., 2009). While the E. festucae ΔpldB strains showed reduced NBT tip staining, indicative of reduced production of superoxide, this synthesis was confined to the mitochondria. Given the more filamentous structure of these structures in hyphal tips of the ΔpldB mutant strains compared to WT, it appears that PldB has a role in organelle structural organisation rather than superoxide production itself. Deletion of components of the Nox complex in E. festucae, A. nidulans, F. graminearum and Sclerotinia sclerotium also results in reduced NBT staining (Kayano et al., 2018, Semighini & Harris, 2008, Tanaka et al., 2006, Wang et al., 2014, Kim et al., 2011). Whether these reported changes in NBT staining are the result of changes in the structure of the mitochondria remains to be determined. Therefore, while there are many phenotypic similarities between ΔpldB and Nox component deletion mutants, a direct connection between PA and the Nox complex could not be demonstrated.
To analyse the effect of PLD deletion on PA concentration and distribution, a PA molecular biosensor was generated using the defined PA-binding domain of Spo20 (Potocký et al., 2014, Nakanishi et al., 2004). Microscopic analysis revealed that the PA biosensor localised to the PM in E. festucae WT and ΔpldA strains, as has been described in S. cerevisiae and in tobacco root hairs (Potocký et al., 2014, Nakanishi et al., 2004). In ΔpldB strains, localization of the PA biosensor was reduced at the PM in hyphal tips and young hyphae and preferentially localised to the cytoplasm and nuclei. The preferential localization to the nuclei is similar to what is observed when cells are treated with the PLD inhibitor n-butanol, resulting in a decrease in the PM/nucleus fluorescence intensity ratio (Potocký et al., 2014, Zeniou-Meyer et al., 2007). Therefore, an increased nuclear as well as cytoplasmic localization likely reflects reduced PLD activity. As the PA biosensor was never detected in nuclei of either the WT or ΔpldA strains, PldA is most likely to be the isozyme responsible for PA synthesis in E. festucae.
Both PLD and PA also have a role in vesicle trafficking, as demonstrated by their roles in endo- and exocytosis in yeast and mammalian cells (Li et al., 2000, Selvy et al., 2011, Donaldson, 2009). The detection of CFW-stained puncta in the ΔpldB strain is indicative of defects in vesicular trafficking (Higuchi et al., 2009, Upadhyay & Shaw, 2008). Using a range of GFP-labelled proteins associated with various steps in vesicle transport, we examined whether the localization of these proteins in ΔpldB was different to WT. However, apart from fimbrin, all the other GFP fusion proteins had the same localization as in WT. Fimbrin localised to actin patches at the endocytic collar, where it is known to crosslink with actin filaments. These filaments coat endocytic vesicles and provide mechanical force for membrane invagination (Riquelme et al., 2018, Aghamohammadzadeh & Ayscough, 2009). In the ΔpldB deletion strains, the Fimbrin-GFP fusion protein, while also detected at the endocytic collar, was found to collect at the hyphal tip more often than in WT hyphae. Fimbrin is known to oscillate between the hyphal tip and the endocytic collar and only concentrates at the endocytic collar in actively growing hyphae (Upadhyay & Shaw, 2008). While this increased tip localization of fimbrin in the ΔpldB strains could be a result of the slower growth rate of the ΔpldB strain it could also be indicative of a key role for PldB in endocytosis, as has been described in mammalian cells (Donaldson, 2009). Interestingly, deletion of fimbrin in A. nidulans, results in swollen hyphae, reduced spore germination and an accumulation of CFW in the cell, defects that are associated with loss of cell polarity (Upadhyay & Shaw, 2008). While fimbrin itself is unlikely to bind PA or interact with PLDs these observations suggest that PA or PLDs are involved with localization of the endocytic machinery and/or establishment of cell polarity.
In conclusion, we have demonstrated that PldB/PLA-7, one of two PLDs found in E. festucae and N. crassa, has an important role in hyphal morphogenesis and development. Crucially, PldB generated synthesis of PA is required for normal polarized hyphal growth, cell fusion and ascospore formation. Additionally, we have shown that PldB is required for expressoria formation and development of a mutualistic symbiotic interaction between E. festucae and L. perenne. Given the changes in lipid membrane structure and composition following the PLD dependent synthesis of PA, it will be of considerable future interest to identify which proteins are recruited to these lipid sites to initiate cell-to-cell communication and cell fusion.
Experimental procedures
Bioinformatic analyses
E. festucae sequences were obtained from the Kentucky Endophyte Database (Schardl et al., 2013). Protein sequences were annotated using InterproScan (Jones et al., 2014) and SMART (Letunic et al., 2015). Domains in the PLDs were annotated based on previous publications (Sciorra et al., 1999, Sciorra et al., 2002, Sung et al., 1997). BLASTp analyses were performed against fungal reference proteins at NCBI. Alignments were generated using MAFFT version 7 (Katoh et al., 2017) and shaded according to percentage identity in Jalview v1. The Ensembl genome database (Zerbino et al., 2018) was used to compare E. festucae PldA and PldB proteins to other PLD sequences by first obtaining a pre-computed protein alignment for the gene family with ID “EGGT00050000001203” and then adding the Epichloë sequences to this alignment using the “profile” method implemented in Muscle v3.8.31 (Edgar, 2004). A maximum likelihood phylogeny was estimated from the combined alignment using RAxML v 8.2.12 (Stamatakis, 2014) and the BLOSUM62 model of amino acid substitution. Box plots were generated using BoxPlotR (http://shiny.chemgrid.org/boxplotr/). All statistical tests are explained in detail in the Methods S1.
Strains and growth conditions
Escherichia coli strains were grown in Lysogeny Broth (LB) (Miller, 1972) or on LB solidified with 1.5% (w/v) agar, supplemented with 100 µg/ml ampicillin or 50 µg/ml kanamycin at 37°C. S. cerevisiae strains were grown in YPDA liquid medium (Green, 2016) or on YPDA supplemented with 2% (w/v) agar overnight or for up to 3 days, respectively, at 30°C. Following transformation, strains were selected on SD synthetic -T/L dropout medium (Green, 2016), and incubated at 30°C until colonies appeared. For yeast two-hybrid assays strains were grown in liquid SD-T/L broth, then plated on SD-T/L/H and -T/L/H/A dropout media (Green, 2016), and were incubated for up to 5 days at 30°C. E. festucae strains were grown on Potato-Dextrose (PD) agar (2.4% PD, 1.5% agar, PDA) supplemented with 150 µg/ml hygromycin B and/or 200 µg/ml geneticin for up to 7 days at 22°C. For chemical complementation E. festucae strains were inoculated on 20 ml PDA and plates overlaid with 7 ml PDA containing 0.005, 0.02, 0.05 and 0.1 mg/ml 3-sn-phosphatidic acid sodium salt (Sigma) respectively, and grown for 7 days. N. crassa strains were cultivated on Vogel’s minimal medium supplemented with 2% sucrose as the carbon source and with 1.5% agar for solid media (Vogel, 1956). For propagation and spore production, strains were grown on agar slants and incubated at 30°C for 5-7 days. Crosses were performed on Westergaard’s medium at 25°C (Westergaard & Mitchell, 1947). For single spore isolations, conidia were plated on BDES medium agar plates (Brockman & de Serres, 1963) containing 200 µg/ml hygromycin. After an overnight incubation, individual colonies were isolated and transferred to agar slants.
All strain details are listed in Table S1.
Plant growth and endophyte inoculation
E. festucae strains were artificially inoculated into L. perenne cv Samson seedlings and plants grown as described previously (Latch & Christensen, 1985, Becker et al., 2015).
DNA isolation, PCR and sequencing
Genomic DNA was purified from freeze-dried E. festucae mycelia as described previously (Byrd et al., 1990). Plasmid DNA was extracted from E. coli strains using the High Pure plasmid isolation kit (Roche, Basel, Switzerland). PCR amplification of short DNA fragments (<3000 bp in length) was conducted using OneTaq® DNA polymerase (New England Biolabs (NEB), Ipswich, USA) and amplification of DNA fragments for cloning purposes was performed with the PhusionTM DNA polymerase (NEB) according to the manufacturer’s instructions. DNA fragments were separated by agarose-gel electrophoresis and purified with the Wizard SV Gel and PCR Clean-Up System (Promega, Madison, USA). Sequence authenticity was confirmed at the Massey Genome Centre. Sequence data was assembled (ClustalW) and analysed using MacVector v14.5.2.
Preparation of constructs
All constructs were designed using MacVector v14.5.2 and synthesised using Gibson assembly (Gibson et al., 2009). Primer sequences are listed in Supporting Information Table S2 and descriptions of construct designs in Methods S2.
Transformation of organisms
Chemically competent E. coli DH5α cells were transformed as described previously (Hanahan, 1983). S. cerevisiae AH109 cells were made competent using lithium acetate and transformed via heat-shock at 42°C for 20 min (Gietz & Woods, 2002). E. festucae protoplasts were generated as described elsewhere (Young et al., 2005), and transformed with 1.2-5 µg of each plasmid (Itoh et al., 1994). Protoplasts were allowed to recover overnight on regeneration (RG, PD with 0.8 M sucrose) medium supplemented with 1.5% agar before being overlaid with RG medium solidified with 0.8% agar and the appropriate antibiotic. For the deletion of pldA a linear 5’pldA-hph-3’pldA replacement cassette was amplified by PCR from pCE131 and transformed into wild-type (WT) protoplasts. For the deletion of pldB, the linear 5’pldB-nptII-3’pldB replacement cassette was amplified by PCR in two fragments from pCE121 with an overlapping region of approx. 600 bp (Rahnama et al., 2016). Protoplasts of the ΔpldB strain were allowed to recover for two days before being overlaid with 20-50 µg/ml hygromycin B.
Protein extraction and western blot analysis
E. festucae strains were grown shaking in PD liquid medium for 3-4 days at 21°C. The washed and snap frozen mycelium was ground to a fine powder and resuspended in 0.2-1 mL of lysis buffer (50 mM Tris-HCL, 100 mM NaCl, 10 mM EDTA, 0.2 mM PMFS, 0.1 mM DTT, 1 µl/mL NP40, 1 µl/mL Protease inhibitor cocktail (Roche)). Cell debris was removed by centrifugation and 15-20 µl of the protein suspension was resolved on a 7 or 10% SDS-polyacrylamide gel. After electrophoresis proteins were transferred to a PVDF membrane. The membrane was incubated with a 1:3,000 dilution of primary rabbit anti-GFP antibody (ab290, Abcam, Cambridge, UK) or a 1:2,000 dilution of primary rabbit anti-mCherry antibody (ab167453, Abcam) overnight. This was followed by incubation with a 1:10,000 dilution of secondary goat anti-rabbit horseradish peroxidase (HRP) antibody (ab6721, Abcam) for one hour. Detection of HRP was performed using the Amersham ECL Western blotting detection reagent (GE Healthcare, Chicago, USA).
RNA extraction and qRT-PCR
RNA was extracted from snap-frozen and ground mycelia using TRIzol (Invitrogen, Carlsbad, USA), and cDNA synthesis performed with the QuantiTect Reverse Transcriptase kit (Qiagen, Venlo, the Netherlands). RT-qPCRs were performed with SYBR Green (Invitrogen) on a LightCycler 480 System (Roche), as described by the manufacturer. Each sample was analysed with two technical replicates. E. festucae translation elongation factor 2 (EF-2, EfM3.021210) and 40S ribosomal protein S22 (S22, EfM3.016650) genes were used as references (Lukito et al., 2015). The gene expression level in the overexpression strains was calculated relative to the expression in WT (Lukito et al., 2015). Primer sequences used for this analysis (BH185-BH188) are listed in Table S2.
Microscopy
For analysis of the hyphal morphology and localization of proteins, E. festucae strains were grown on a microscopic slide covered with 1.5% H2O agar for 5-7 days and analysed with an Olympus IX83 inverted fluorescence microscope. Cell walls were stained using 3 µl of a 3 mg/ml Calcofluor White (CFW) stock solution and nuclei with 3 µl of a 0.071 mM 4’,6-diamidino-2-phenylindole (DAPI) solution. For quantification of fusion events, colonies were incubated for 14 days before analysis. For chemical complementation, E. festucae strains were inoculated on 20 ml 1.5% H2O agar plates with a microscopic slide, overlaid with 7 ml 1.5% H2O agar containing 0.005 mg/ml, 0.02 mg/ml, 0.05 mg/ml, 0.1 mg/ml 3-sn-Phosphatidic acid sodium salt (Sigma) respectively, and grown for 7 days. Quantitative spore germination and fusion assays in N. crassa were conducted as described in (Fleißner et al., 2009). To observe hyphal fusion, strains were grown on Vogel’s minimal medium overnight at 30°C. Agar blocks of 1 cm2 were cut from the cultures and analysed by light microscopy.
For transmission electron microscopy (TEM), infected plant tissue was fixed and embedded as described by (Spiers & Hopcroft, 1993). Samples were analysed with a FEI Tecnai G2 Biotwin TEM. In addition, sections were stained with 0.05% toluidine blue in phosphate buffer and heat-fixed at approx. 100°C for 10 sec. These sections were examined with a Zeiss Axiophot Microscope with Differential Interference Contrast (DIC) Optics and Colour CCD camera.
To evaluate hyphal growth in planta, infected pseudostem samples were stained with aniline blue diammonium salt (Sigma-Aldrich, St. Louis, USA) and wheat germ agglutinin conjugated to AlexaFluor®488 (WGA-AF488, Molecular Probes/Invitrogen) (Becker et al., 2016, Becker et al., 2018), and examined with a Leica SP5DM6000B confocal microscope (488 nm argon and 561 nm DPSS laser, 40× oil immersion objective, NA = 1.3) (Leica Microsystems, Wetzlar, Germany).
Superoxide detection
For nitroblue tetrazolium (NBT, Sigma-Aldrich) staining, colonies were grown and stained as reported previously (Tanaka et al., 2006). Colonies were photographed and imaged with a Zeiss Axiophot microscope as described. To quantify hyphal staining patterns, two colonies of each strain of interest were stained and three times approx. 50 hyphae evaluated for each.
Superoxide was detected using the ROS-ID® Superoxide (ENZO LifeSciences, Farmingdale, USA) kit. The stock was diluted 1:30 (0.16 mM) in H2O and 3 µl was added to 7-day old colonies, followed by a 10 min incubation in the dark. For mitochondria identification a stock of Mitochondrial Staining Reagent-Green (ab176830, Cytopainter, Abcam) was diluted 1:500 in H2O and 3 µl added to the colonies. After 20 min incubation in the dark the ROS-ID ® Superoxide probe was added.
Yeast two-hybrid
All plasmids used in yeast two-hybrid assays are listed in Table S1. Yeast strain AH109 was transformed as described above with each assay repeated twice as previously described (Green, 2016).
Author contributions
BH, CJE, CHM, AF and BS planned and designed the research. BH, CJE, KAG, DW, UB performed the experiments. BH, CJE, KAG, DW, MSS, CHM, AF and BS analyzed the results. CHM, AF and BS supervised the project. BH, CHM, AF and BS wrote the manuscript.
Acknowledgements
This research was supported by grants from the Tertiary Education Commission to the Bio-Protection Research Centre, the Royal Society of New Zealand Marsden Fund (MAU1301) and by Massey University. BH was supported by a Massey University PhD studentship and BS by an Alexander von Humboldt Research Award. We thank Niki Murray and Pani Vijayan (Manawatu Microscopy and Imaging Centre) for support with microscopy, Arvina Ram and Alyesha Candy for technical assistance and Christopher Schardl for provision of Epichloë genome sequences.