ER membrane receptor FgGET2 plays critical roles in hyphal growth, asexual development, stress responses, and pathogenicity of Fusarium graminearum

Abstract

Fusarium graminearum is the major causal agent of Fusarium head blight (FHB), a disease that considerably decreases wheat yield and quality. GET2 is an endoplasmic reticulum (ER) membrane receptor protein in the guided entry of tail-anchored (TA) proteins (GET) pathway, which is the most widely studied post-translational pathway for targeting and inserting TA proteins into the ER. Previous studies indicated that GET2 and its homologs play diverse biological roles in different organisms, including fungi, mammals, and plants. In this study, integrated biochemical, microbiological, and molecular genetic approaches were used to investigate the roles of GET2 in F. graminearum (FgGET2). FgGET2 has a conserved primary structure and is localized in the ER. In the Saccharomyces cerevisiae ∆ScGET2 mutant, the ectopic expression of FgGET2 restored normal growth. Deleting FgGET2 in F. graminearum had detrimental effects on vegetative growth polarity, vacuolar morphology, and conidial production, morphology, and germination. In addition, the lack of FgGET2 expression disrupted responses to different stress conditions (treatments with environmental stressors, e.g., metals, fungicides, the ER stressor, and DNA replication and damage stressors) to varying degrees. Furthermore, deleting FgGET2 in F. graminearum resulted in decreased pathogenicity on wheat spikes but increased deoxynivalenol production in wheat spikes and liquid medium. Re-introducing a functional FgGET2 into ∆FgGET2 recovered the wild-type phenotype. Collectively, these findings underscore the critical roles of FgGET2 in influencing diverse cellular and biological processes essential for F. graminearum growth and pathogenicity.

Background

Fusarium head blight (FHB) of wheat is a devastating disease that is primarily caused by the filamentous fungus Fusarium graminearum (Xu and Nicholson 2009; Dean et al. 2012). This disease decreases grain yield but also threatens human and animal health because of the presence of mycotoxins, such as deoxynivalenol (DON), in contaminated grains (Vidal et al. 2018; Chen et al. 2019; Duan et al. 2020). There is currently no wheat germplasm immune to FHB; the wheat germplasm most tolerant to FHB is only able to delay the spread of the pathogenic fungus within spikes (Li et al. 2019; Su et al. 2019). Hence, the application of synthetic chemical fungicides is still the major strategy for controlling FHB. However, because there are only a few effective fungicides with different modes of action, their continued use may lead to F. graminearum developing increased tolerance or even resistance, making them less effective for controlling FHB. Therefore, clarifying the molecular mechanisms underlying F. graminearum pathogenicity is essential for efficient FHB management.

The guided entry of tail-anchored (TA) proteins (GET) pathway in yeast (Saccharomyces cerevisiae) is the most extensively studied post-translational pathway for targeting and inserting TA proteins, a unique class of integral membrane proteins, into the endoplasmic reticulum (ER) (Chio et al. 2017; Mateja et al. 2018). In this pathway, a pre-targeting complex receives the TA protein from the ribosome and transfers it to the homodimeric ATPase ScGET3, resulting in the formation of a targeting complex (TA-ScGET3) in the cytoplasm (Chang et al. 2010; Suloway et al. 2012). ScGET2, an ER membrane receptor protein, may capture this TA-ScGET3 complex in the cytoplasm (Wang et al. 2011; McDowell et al. 2020). Subsequently, ScGET2 combines with another ER membrane receptor protein, ScGET1, to form a receptor complex that facilitates the terminal integration of TA proteins into the ER membrane (Yamamoto and Sakisak 2012; Zalisko et al. 2017; Heo et al. 2023).

In mammals, the transmembrane domain recognition complex (TRC) targeting pathway is homologous to the GET pathway, with the calcium-modulating cyclophilin ligand protein-encoding gene (CAML) functionally equivalent to ScGET2 (Vilardi et al. 2011; Yamamoto and Sakisaka 2012). However, because of the diversity in their amino acid (aa) sequences, CAML and ScGET2 were not initially considered homologs (Yamamoto and Sakisaka 2012). It was only after an analysis using a position-specific iterated (PSI)-basic local alignment search tool (BLAST) that CAML and ScGET2 were revealed to have a common evolutionary origin. Their homologs are relatively ubiquitous among eukaryotes (Borgese 2020). Subsequently, AtGET2 in Arabidopsis thaliana was identified on the basis of its function and physical interactions with other GET pathway components, providing a new aa sequence that is better for identifying homologs in different eukaryotic groups, including ascomycetes (Asseck et al. 2021).

GET2 proteins in different eukaryotes exhibit equally low sequence conservation, but they share the same overall topology, including a cytosolic domain (CD) at the N terminus, three transmembrane domains (TMDs), and a luminal C terminus (Carvalho et al. 2019; Asseck et al. 2021). Most importantly, near the N terminus, there is a cluster of positively charged aa (at least four consecutive arginine or lysine residues) that is highly conserved among eukaryotes (Carvalho et al. 2019; Asseck et al. 2021). This motif is indispensable for the binding to ScGET3 (Stefer et al. 2011) and its homologs in both mammals (Yamamoto and Sakisaka 2012) and plants (Asseck et al. 2021). An earlier analysis of GET2 sequences in ascomycetes detected the motif R/KER/K in the N-terminal region (Asseck et al. 2021).

In S. cerevisiae, the ScGET2 knock-out (KO) mutant (∆ScGET2) exhibited pleiotropic defects, including improper protein retention in the ER, decreased cell growth under various stress conditions (Schuldiner et al. 2008; Kiktev et al. 2012), abnormal mitochondrial morphology and mitophagy (Dimmer et al. 2002; Okreglak and Walter 2014; Onishi et al. 2018), and defective meiotic nuclear division (Enyenihi and Saunders 2003; Auld et al. 2006). In mammals, CAML has been identified as a regulator of numerous signal transduction systems (Yamamoto and Sakisaka 2012). Its KO mutants exhibited early embryonic lethality (Tran et al. 2003). Moreover, a lack of CAML expression in the inner ear led to a severe loss of cochlear hair cells and complete deafness (Bryda et al. 2012). In A. thaliana, AtGET2 KO mutant seedlings showed shorter root hairs than wild-type (WT) seedlings (Asseck et al. 2021). Furthermore, PfGET2 in the human malaria parasite Plasmodium falciparum can rescue the CuSO₄-sensitive phenotype of the ΔScGET2 mutant (Kumar et al. 2021). Despite these findings, the biological role of GET2 in filamentous fungi remains unclear.

In this study, we identified the physiological consequences of losing a functional GET2 in F. graminearum. We determined that FgGET2 is required for various processes, including vegetative growth, vacuole morphogenesis, multiple stress responses, asexual development, and pathogenicity. These findings may be useful for developing effective strategies for minimizing the threat of FHB.

Results

Identification of GET2 in F. graminearum

To identify the FgGET2 gene, the deduced aa sequences of ScGET2 from S. cerevisiae (Protein ID NP_011006.1), AtGET2 from A. thaliana (NP_567900.1), CAML from Homo sapiens (NP_001736.1), and PfGET2 from P. falciparum (XP_001349661.1) were used for BLAST searches for their homologs in F. graminearum. These searches identified FGSG_06264 (XP_011324914.1) as the sole candidate. Because of the diversity and low sequence similarity in GET2 family sequences across kingdoms, to ensure that authentic homologous genes were not overlooked, proteins in the F. graminearum proteome that met the following two conditions (Fig. 1a) were identified: (i) presence of the R/KER/KR motif, which is a representative conserved motif in ascomycetes (Asseck et al. 2021), within 50 aa of the N terminus and (ii) presence of three TMDs. Employing this strategy, FGSG_06264 was also retrieved as the sole hit (Fig. 1a).

Fig. 1

Identification of GET2 in F. graminearum. a Flowchart illustrating the bioinformatics-based prediction of FgGET2 in the F. graminearum proteome. b FgGET2 domain structure and positively charged motifs characterized by four consecutive arginine or lysine residues in the N terminus of ScGET2, AtGET2, CAML, and PfGET2. Arginine (R) and lysine (K) are highlighted in gray. aa, amino acid; TMD, transmembrane domain

The full-length FgGET2 sequence consists of 1067 bp, which includes a 59 bp intron. It encodes a polypeptide comprising 335 aa, with a highly conserved protein domain structure similar to that of other GET2 proteins (Fig. 1b). Although the aa sequence similarity between FgGET2 and ScGET2, AtGET2, PfGET2, and CAML was relatively low (i.e., sequence identities of 16, 15, 8, and 12%, respectively), a conserved group of positively charged aa residues near the N terminus was detected in all of these proteins (Fig. 1b).

Subcellular localization of FgGET2 in F. graminearum

We determined the subcellular localization of FgGET2 using the F. graminearum strain overexpressing the FgGET2-eGFP fusion protein. The green fluorescence of FgGET2-eGFP was detected in conidia (Fig. 2a), conidia germinated in liquid yeast extract peptone dextrose (YEPD) medium for 6 h (Fig. 2b), and hyphae grown in liquid YEPD medium for 24 h (Fig. 2c). Notably, in hyphae, FgGET2-eGFP was co-localized with the ER-Tracker Red signal (Fig. 2c), indicating that FgGET2 is localized to the ER, similar to other GET2 proteins.

Fig. 2

Subcellular localization of FgGET2. a Green fluorescence in conidia. b Green fluorescence in germinating conidia. c Co-localization of FgGET2 and ER-Tracker Red in vegetative hyphae. Bar, 10 μm; DIC, differential interference contrast

FgGET2 restores the defective growth of the yeast ∆ScGET2 mutant

An earlier study revealed phenotypic changes in the ∆ScGET2 mutant under various stress conditions (Schuldiner et al. 2008). To determine whether FgGET2 can compensate for the deletion of GET2 in ∆ScGET2, the expression vector pYES2-FgGET2 containing the full-length FgGET2 cDNA sequence was constructed and inserted into ∆ScGET2. As controls, ∆ScGET2 was transformed with pYES2-ScGET2 (positive control) and the empty pYES2 vector (negative control). As expected, WT yeast cells grew well on YEPD medium regardless of the presence of different stressors, whereas ∆ScGET2 cell growth was significantly compromised in the presence of hydroxyurea, tunicamycin, hygromycin, or CuSO₄ and at 39°C (i.e., heat stress). Notably, the defective growth of ∆ScGET2 was resolved by the transformation of mutant cells with either pYES2-ScGET2 or pYES2-FgGET2, but not with the empty pYES2 vector (Fig. 3). Accordingly, ScGET2 and FgGET2 may have similar functions.

Fig. 3

FgGET2 rescues the growth defects of the ∆ScGET2 mutant. ∆ScGET2 was transformed with pYES2-ScGET2 and pYES2-FgGET2. WT strain BY4741 and ∆ScGET2 transformed with empty pYES2 were used as controls. For each strain, cell suspensions (different dilutions) were spotted on YEPD medium supplemented with the indicated stress agents: 200 mM hygromycin, 2 μg/mL tunicamycin, and 200 μg/mL hydroxyurea at 30°C; 3 mM CuSO₄ at 37°C. For the heat sensitivity assay, yeast strains were grown at 39°C. The growth of each strain was assessed after a 3-day incubation

Construction of ∆FgGET2 and C-FgGET2 mutants

To elucidate the biological roles of FgGET2, we generated the ΔFgGET2 KO mutant using a homologous recombination strategy (Fig. 4a). Briefly, the FgGET2-pRF-HU2 recombinant plasmid (hygromycin resistance) was constructed (Fig. 4b) and transformed into the WT strain. The FgGET2 coding sequence was replaced with the hygromycin phosphotransferase (HPH)-encoding gene via homologous recombination (Fig. 4a). Hygromycin-resistant transformants were confirmed by a polymerase chain reaction (PCR) amplification (Fig. 4c).

Fig. 4

Production of ΔFgGET2 and C-FgGET2 mutants. a Constructs and gene structures for disrupting FgGET2. The upstream (Up) and downstream (Down) flanking sequences were amplified by PCR using DNA from the wild-type (WT) strain and Up-F/R and D-F/R primer pairs, respectively, and then cloned into the pRF-HU2 vector to construct the FgGET2-pRF-HU2 plasmid. SacI, ApaI, SpeI, and HindIII indicate the restriction enzymes used. ΔFgGET2 mutants were generated by replacing FgGET2 in WT with the hygromycin phosphotransferase (HPH)-encoding gene. b Validation of FgGET2-pRF-HU2 using primer pairs Up-F/R, D-F/R, P1/P2, and P3/P4. c PCR verification of ΔFgGET2 using primer pairs M-Up-F/R and M-D-F/R. d Reverse transcription PCR-based verification of the expression of a hygromycin resistance gene, neomycin resistance gene, and FgGET2 in ΔFgGET2, WT, and C-FgGET2 using primers HPH-F/R(H), NEO-F/R (N), and M-FgGET2-F/R (M), respectively

To obtain the complemented mutant, the JM45-FgGET2 plasmid (neomycin resistance) carrying a 2.4-kb region containing the FgGET2 coding sequence and its promoter sequence was constructed and inserted into the ∆FgGET2 mutant. The complemented mutant (designated C-FgGET2), which was resistant to both hygromycin and neomycin, was identified on the basis of reverse transcription (RT)-PCR (Fig. 4d). According to the RT-PCR results, FgGET2 was not expressed in ∆FgGET2, but was expressed normally in C-FgGET2 as in the WT control (Fig. 4d). Hence, FgGET2 was correctly removed from the genome of ∆FgGET2 and was successfully re-introduced in C-FgGET2.

FgGET2 plays crucial roles in vegetative polar growth and vacuole fusion

During the F. graminearum life cycle, hyphal growth is critical for survival (Riquelme et al. 2018). To assess the importance of FgGET2 for vegetative growth, WT, ∆FgGET2, and C-FgGET2 strains were cultured on potato dextrose agar (PDA) and modified synthetischer nährstoffarmer agar (mSNA) media. After 4 days, the growth rate of ∆FgGET2 was significantly lower than that of the other two strains on both media (Fig. 5a, b). Microscopic examinations revealed ∆FgGET2 produced crooked and narrow hyphae with multiple branches at the hyphal tips (Fig. 5c), suggesting that FgGET2 is important for maintaining hyphal tip polarity.

Fig. 5

FgGET2 is required for vegetative growth and vacuole morphology. a WT, ΔFgGET2, and C-FgGET2 mycelial growth on PDA and mSNA media incubated for 4 days at 25°C in darkness. b Colony diameters on PDA and mSNA media. Data were analyzed by Student’s t-test (*** P < 0.001). Means and standard deviations were calculated for three replicates. c WT, ΔFgGET2, and C-FgGET2 hyphal tip and branching patterns on PDA medium. Bar, 20 µm. d Vacuole structures in WT and ΔFgGET2 hyphae as observed using a transmission electron microscope. Bar, 1 µm. e WT and ΔFgGET2 hyphae stained with a vacuole-tracking dye (CMAC) and examined using a fluorescence microscope. DIC, differential interference contrast. Bar, 10 µm

An examination using a transmission electron microscope detected small, fragmented vacuoles with numerous irregular spheroids in more than 90% of ∆FgGET2 hyphal cells, whereas approximately 70% of WT hyphal cells contained a single large vacuole (Fig. 5d). The presence of fragmented vacuoles was confirmed by staining hyphae with the vacuole marker 7-amino-4-chloromethylcoumarin (CMAC) (Fig. 5e). These results suggest FgGET2 may also be important for vacuole fusion in F. graminearum.

FgGET2 affects conidia production, morphology, and germination

Conidia are critical for the spread of disease in the field (Osborne and Stein 2007). To compare the conidia production of WT, ∆FgGET2, and C-FgGET2 strains, liquid carboxymethylcellulose (CMC) medium was inoculated with equal amounts of fresh mycelia from each strain. After 4 days, ∆FgGET2 mycelia were relatively thick and aggregated in the culture, whereas WT and C-FgGET2 mycelia were uniformly dispersed (Fig. 6a). WT and C-FgGET2 conidia formed on phialides. In contrast, phialides were rarely observed in ∆FgGET2 (Fig. 6a). In addition, foot cells were present in WT and C-FgGET2 conidia but not in ∆FgGET2 conidia (Fig. 6a). The number of conidia produced by ΔFgGET2 decreased by more than 90% compared to that of WT and C-FgGET2 (Fig. 6b). A microscopic examination of 300 conidia per strain showed that ∆FgGET2 conidia were shorter (Fig. 6c) and had fewer septa (Fig. 6e) than WT and C-FgGET2 conidia, but there was no significant difference in conidial width (Fig. 6d). Considered together, these results showed that FgGET2 affects conidial production and morphology in F. graminearum.

Fig. 6

Comparison of WT, ∆FgGET2, and C-FgGET2 conidial production and germination. a Mycelia, phialides (indicated by white arrows), and conidial morphology in CMC liquid medium. Bar, 20 μm. b Quantification of conidia produced by WT, ∆FgGET2, and C-FgGET2 strains. Comparisons of conidial length (c) and width (d), with at least 300 conidia examined per strain. Data were analyzed using Student’s t-test (*** P < 0.001). Means and standard deviations were calculated for three replicates. e Percentage of conidia with different numbers of septa. A total of 300 conidia were examined per strain. f Morphology of germinated conidia at 3, 6, and 9 h in YEPD medium. Bar, 20 μm. g Percentage of conidia with different numbers of germ tubes at 6 and 9 h after germination. A total of 300 germinated conidia were examined per strain at each time point

To determine whether FgGET2 contributes to conidial germination, liquid YEPD medium was inoculated with conidia (Fig. 6f). Although ΔFgGET2 conidia germinated, the germination was significantly delayed. At 6 h, only 9% of ΔFgGET2 conidia produced three or more germ tubes, whereas more than 40% of WT and C-FgGET2 conidia had at least three germ tubes. At 9 h, 25% of ΔFgGET2 conidia produced three or more germ tubes, which was considerably lower than the corresponding percentage of WT and C-FgGET2 conidia (> 80%) (Fig. 6g). These results demonstrate that FgGET2 influences conidial germination and germ tube polarity in F. graminearum.

FgGET2 is essential for F. graminearum stress responses

Sensitivity to environmental stresses is a critical factor modulating F. graminearum hyphal development and pathogenicity (Chong et al. 2020). Moreover, impaired vacuole fusion affects stress responses in fungi (Yu et al. 2014; Liu et al. 2017). Therefore, ΔFgGET2 responses to several environmental stresses were examined (Fig. 7a, e). After 4 days on PDA medium supplemented with osmotic (1 M NaCl and 1 M D-Sorbitol) and oxidative (0.05% H₂O₂) stressors, the growth inhibition rate of ΔFgGET2 was significantly lower than that of WT or C-FgGET2. Under cell wall stress conditions, ΔFgGET2 showed increased sensitivity to 0.5 mM Congo Red (CR), but not to 0.025% sodium dodecyl sulfate. Under metal stress conditions, the sensitivity of ΔFgGET2 mycelia to 5 mM Cu²⁺ and 10 mM Zn²⁺ was unaffected. However, ΔFgGET2 exhibited increased tolerance to 0.4 M Mg²⁺ and increased sensitivity to 10 mM Fe²⁺ (Fig. 7b, f). Furthermore, in a fungicide sensitivity assay, ΔFgGET2 exhibited increased sensitivity to 0.6 μg/mL pyraclostrobin, but not to 0.4 μg/mL carbendazim or 5 μg/mL tebuconazole (Fig. 7c, g). These results suggest that FgGET2 is involved in regulating F. graminearum responses to osmotic and oxidative stresses to maintain cell wall integrity and cellular metal homeostasis, while also enhancing fungicide resistance.

Fig. 7

WT, ΔFgGET2, and C-FgGET2 vegetative growth under different stress conditions. Strains were grown on PDA medium supplemented with a 1 M NaCl and 1 M D-Sorbitol, 0.05% H₂O₂, 0.5 mM Congo Red (CR), and 0.025% sodium dodecyl sulfate (SDS), b 0.4 M Mg²⁺, 5 mM Cu²⁺, 10 mM Zn²⁺, and 10 mM Fe²⁺, c 0.4 μg/mL carbendazim, 5 μg/mL tebuconazole, and 0.6 μg/mL pyraclostrobin, d 5 mM DTT, 100 μg/mL methyl methanesulfonate (MMS), and 100 μg/mL 6-Azauracil. All strains were grown at 25°C for 4 days. e, f, g, and h Inhibition of mycelial growth for each strain exposed to individual stresses. Data were analyzed using Student’s t-test (** P < 0.01; *** P < 0.001). Means and standard deviations were calculated for three replicates

Deleting ScGET2 can lead to defects in DNA replication or DNA damage in S. cerevisiae (Zewail et al. 2003). Therefore, we examined the response of ΔFgGET2 to 100 μg/mL methyl methanesulfonate (MMS) and 100 μg/mL 6-Azauracil, which are known to damage DNA. As expected, ΔFgGET2 exhibited increased sensitivity to these DNA-damaging agents (Fig. 7d, h). In yeast, a loss-of-function deletion to ScGET2 reportedly results in the cytoplasmic aggregation of unfolded ER-destined TA proteins, leading to ER stress (Schuldiner et al. 2008). To investigate whether the deletion of FgGET2 in F. graminearum also triggers ER stress, we analyzed ΔFgGET2 grown on PDA medium containing the ER stressor dithiothreitol (DTT). We observed that ΔFgGET2 exhibited increased sensitivity to 5 mM DTT.

FgGET2 affects fungal pathogenicity and DON biosynthesis

To determine the effect of FgGET2 on pathogenicity, flowering wheat heads were inoculated with 1 × 10³ conidia. At 12 days post-inoculation (dpi), obvious head blight symptoms were detected throughout the spikes inoculated with WT or C-FgGET2, whereas disease symptoms in the spikes inoculated with ΔFgGET2 were restricted to the spikelets (Fig. 8a, b). At 2 dpi, the fungal biomass of the spikes inoculated with ΔFgGET2 was significantly lower than that of the spikes inoculated with WT and C-FgGET2 (Fig. 8c). To further analyze the effect of FgGET2 on the penetration behavior of F. graminearum, ΔFgGET2, WT, and C-FgGET2 were grown on cellophane membranes placed on PDA medium. The ability of ΔFgGET2 to penetrate cellophane was weaker than that of WT or C-FgGET2 (Fig. 8d, e).

Fig. 8

Effect of FgGET2 on F. graminearum pathogenicity and DON production. a Flowering wheat spikes were inoculated with WT, ΔFgGET2, and C-FgGET2 conidial suspensions. Spikes, seeds, and rachises were photographed at 12 dpi. b Numbers of infected and bleached spikelets at 12 dpi. c Fungal biomass was determined on the basis of the relative expression of FgTUB2 in spikelets at 2 dpi. d Penetration of a cellophane membrane by WT, ΔFgGET2, and C-FgGET2. The indicated strains were grown on top of cellophane membranes on PDA medium for 4 days as a control (‘Before’ treatment). To assess mycelial growth on the medium, which indicates the cellophane membrane was penetrated, the cellophane with the fungal colony was removed on day 3 and the medium was incubated for an additional day (‘After’ treatment). e Percentages of penetration area for WT, ΔFgGET2, and C-FgGET2 on cellophane membranes. f DON production in wheat spikes infected with conidial suspensions of the indicated strains at 8 dpi. g Relative expression of FgTUB2 in the indicated strains in wheat heads at 8 dpi. h DON concentration in the supernatants obtained after a 24-h incubation of the indicated strains in liquid medium. i Expression of TRI genes in WT and ΔFgGET2 hyphae cultured in liquid medium for 24 h. j Hyphae bulbous structures of the indicated strains after a 24-h incubation in liquid medium. Bar, 20 μm. Data were analyzed using Student’s t-test (** P < 0.01; *** P < 0.001). Means and standard deviations were calculated for three replicates

To explore how FgGET2 influences mycotoxin production, the amount of DON in infected wheat spikes was quantified at 8 dpi. Compared with the spikes inoculated with WT and C-FgGET2, the spikes inoculated with ΔFgGET2 had significantly less DON (16 and 23% less, respectively) (Fig. 8f) and lower fungal biomass (44 and 47% lower, respectively) (Fig. 8g). Thus, the DON content:fungal biomass ratio was higher for the samples inoculated with ΔFgGET2 than for the samples inoculated with WT or C-FgGET2. Hence, the deletion of FgGET2 enhanced the accumulation of DON in F. graminearum on wheat spikes. To confirm this finding, the accumulation of DON was subsequently assayed in liquid medium. As expected, compared with WT, ΔFgGET2 produced approximately threefold more DON (Fig. 8h). In accordance with this finding, the expression levels of DON biosynthesis-related genes (TRI5, TRI6, and TRI101) were up-regulated in ΔFgGET2 (Fig. 8i). A previous study showed that DON production is involved in the formation of intercalary swollen hyphal compartments (Jiang et al. 2016). Therefore, ΔFgGET2, WT, and C-FgGET2 hyphae were compared in terms of morphology under DON-inducing conditions. Unexpectedly, there were no obvious differences in hyphal structures and bulbous numbers (Fig. 8j). These results indicate that the increased DON production in liquid medium was achieved by triggering the expression of TRI genes rather than by altering the structure of DON-producing hyphae. On the basis of these findings, we conclude that FgGET2 is essential for the penetration of wheat plants and subsequent spread as well as the production of DON in wheat spikes and liquid medium.

Discussion

The ER membrane receptor ScGET2 in the GET pathway in yeast and CAML in the TRC pathway of vertebrates mediate the final step of the process that integrates TA proteins into the ER membrane (Wang et al. 2011). Compared with GET2 in yeast and mammals, GET2 in the phytopathogenic fungus F. graminearum has not been thoroughly investigated in terms of its biological functions. In this study, we identified and functionally characterized FgGET2 in F. graminearum.

Although the FgGET2 aa sequence differs considerably from GET2 sequences in other organisms, it contains a highly conserved protein structure (CD–three TMDs–luminal C terminus), with a cluster of positively charged aa residues near the N terminus (Fig. 1b). In S. cerevisiae, the loss of a functional ScGET2 resulted in a lack of stress tolerance; however, earlier research indicated the simultaneous expression of yeast ScGET1 with either AtGET2 or CAML can weakly restore the viability of a mutant in which both ScGET1 and ScGET2 are knocked out (Vilardi et al. 2011; Asseck et al. 2021). In addition, PfGET2 of P. falciparum can rescue the CuSO₄-sensitivity of the ΔScGET2 strain (Kumar et al. 2021). In the current study, we determined that, as expected, FgGET2 can restore the mutant phenotype of ΔScGET2 (Fig. 3). Moreover, similar to other reported GET2 proteins, FgGET2 is located in the ER (Fig. 2b), providing further evidence that FgGET2 is indeed a GET2 homolog in F. graminearum.

In S. cerevisiae, ScGET2 regulates vegetative growth (Smith et al. 1996; Winzeler et al. 1999), while also mediating responses to diverse abiotic stressors, including cell wall-damaging agents (Hughes et al. 2000), oxidative stress-inducing compounds (Helsen et al; 2020), heavy metals (Ruotolo et al. 2008; Schuldiner et al. 2008), and DNA-damaging compounds (Zewail et al. 2003). ScGET2 can recruit the TA-ScGET3 complex from the cytosol to the ER (McDowell et al. 2020). The absence of ScGET2 leads to the improper insertion of multiple TA proteins that mediate a wide range of vital cellular activities (Borgese and Fasana 2011), suggesting that the pleiotropic effects of knocking out ScGET2 might be a secondary consequence of abnormally localized TA proteins (Schuldiner et al. 2008). Promptly unfolded TA proteins aggregate outside the ER and disrupt the response to unfolded proteins. In earlier studies, the ΔScGET2 mutant exhibited heightened sensitivity to ER stress inducers, including DTT and tunicamycin (Schuldiner et al. 2008; Jonikas et al. 2009). In the present study, defective functions in F. graminearum caused by the lack of FgGET2 expression adversely altered vegetative growth (Fig. 5), asexual development (Fig. 6), and responses to multiple abiotic stresses (Fig. 7). Similarly, the sensitivity of F. graminearum to DTT increased after FgGET2 was knocked out (Fig. 7d). These findings also suggest that the pleiotropic effects of knocking out FgGET2 may be associated with TA protein mislocalization, but this will need to be experimentally verified.

F. graminearum pathogenicity depends on a complex network regulated by multiple factors. In ΔFgGET2, impaired hyphal growth, delayed conidial germination, and altered stress sensitivity may influence pathogenicity. On flowering wheat spikes, FgGET2 expression was essential for hyphal growth in the early infection stages (Fig. 8c) and for the spread through the rachis (Fig. 8a). An analysis of the penetration of a cellophane membrane revealed the importance of FgGET2 (Fig. 8d, e). In addition, DON is another key pathogenicity factor in F. graminearum (Jiang et al. 2016; Chen et al. 2019). Although the total amount of DON in ΔFgGET2 decreased on spikes (Fig. 8f), intriguingly, DON production by ΔFgGET2 mycelia tended to increase on wheat spikes (Fig. 8f, g) and in liquid medium (Fig. 8h), which may be a compensatory mechanism to counterbalance the decreased growth rate and other defects, thereby enabling F. graminearum to spread efficiently. Currently, the application of chemical fungicides is still one of the main approaches for controlling FHB. Notably, ΔFgGET2 exhibited increased sensitivity to pyraclostrobin (Fig. 7c, g), suggesting that FgGET2 may be a potential molecular target for increasing the efficacy of this fungicide.

Conclusions

In summary, our results demonstrate that FgGET2 encodes a key component of the GET pathway and regulates vegetative growth, asexual development, abiotic stress responses, and pathogenicity in F. graminearum.

Methods

Fungal strains and growth conditions

F. graminearum isolate DAOM180378 (Canadian Fungal Culture Collection, AAFC, 270 Ottawa, ON, Canada) is highly virulent in wheat. The colony morphology and growth rate of WT and mutant strains were determined on PDA (Aobox, Beijing, China) and mSNA (1 g KH₂PO4, 1 g KNO₃, 0.5 g MgSO₄, 0.5 g KCl, 1 g glucose, 1 g sucrose, and 20 g agar per liter) media at 25℃ for 4 days in darkness. Mycelial radial growth was estimated as described by Qi et al. (2012).

To determine the sensitivity of F. graminearum to stressors, 5-mm mycelial plugs taken from the edge of a 4-day-old colony were placed on PDA medium supplemented with different stressors (concentrations indicated in figure legends) in Petri dishes. Mycelial radial growth inhibition (RGI) was calculated using the following formula: RGI (%) = [(C−N)/(C−5)] × 100 (Jiang et al. 2011), where C and N represent the colony diameter of the control and treatment, respectively. The experiment was repeated three times, with six Petri dishes per treatment. For the conidiation assay, 20 mg fresh mycelium was used to inoculate 30 mL CMC medium (1 g NH₄NO₃, 1 g KH₂PO₄, 0.5 g MgSO₄‧7H₂O, 1 g yeast extract, and 15 g CMC per liter) in a 50 mL flask (Cappellini and Peterson 1965). The inoculated medium was incubated on a shaker (180 rpm) for 5 days, after which the number of conidia in the flask was determined using a hemocytometer. Conidial germination was assessed using freshly harvested conidia cultured in liquid YEPD medium (20 g peptone, 10 g yeast extract, and 20 g dextrose per liter).

Nucleic acid extraction and PCR

Genomic DNA and total RNA were extracted using Plant genomic DNA Mini Kit and Plant genomic RNA Mini Kit (Biofit, Chengdu, China), respectively. The extracted RNA (1 μg) was reverse transcribed to cDNA using a HiScript III 1st Strand cDNA Synthesis Kit (Vazyme, Nanjing, China). Phanta Max Super-Fidelity DNA Polymerase (Vazyme) was used for PCR and RT-PCR.

To measure the relative fungal biomass in wheat spikelets, the expression of the β-tubulin gene FgTUB2 (FGSG_09530; primer pair TUB2-F/R) was estimated on the basis of quantitative RT-PCR (qRT-PCR), with a wheat glyceraldehyde-3-phosphate dehydrogenase gene GAPDH (Ta.66461; primer pair w-GAPDH-F/R) selected as the reference control for normalizing expression levels. To analyze TRI gene expression, RNA was extracted from hyphae in the DON induction liquid medium for a qRT-PCR analysis using gene-specific primer pairs to amplify the trichothecene synthase genes TRI5 and TRI6 (TRI5-F/R and TRI6-F/R, respectively) and the trichothecene acetyltransferase gene TRI101 (TRI101-F/R). FgTUB2 served as the reference gene. The qRT-PCR analysis was performed using the MyiQ Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA) and the Taq Pro Universal SYBR qPCR Master Mix (Vazyme) according to the manufacturer’s instructions. Transcript levels were calculated using the 2^−ΔΔCt method (Livak et al. 2001), with three biological replicates per treatment. The primers used in this study are listed in Additional file 1: Table S1.

Fungal transformation

∆FgGET2 mutants were produced via Agrobacterium tumefaciens-mediated transformation (Maier et al. 2005). To generate the gene replacement construct, the 598 bp upstream and 454 bp downstream flanking sequences were amplified by PCR using F. graminearum genomic DNA and Up-F/R and D-F/R primer pairs, respectively. The amplified fragments were cloned into the pRF-HU2 vector using a ClonExpress II One Step Cloning Kit (Vazyme). The resulting recombinant plasmid FgGET2-pRF-HU2 was analyzed using primer pairs P1/P2 and P3/P4 to verify the accuracy of the inserted sequence. The FgGET2 gene was replaced by the HPH gene using a homologous recombination strategy (Fig. 4a). Hygromycin B (Calbiochem, La Jolla, CA, USA) was added to a final concentration of 100 mg/mL to screen for transformants. Primer pairs M-Up-F/R and M-D-F/R were used to examine the construct in ∆FgGET2 by PCR. Only transformants that underwent homologous recombination in the target region produced the expected PCR products. At least three ∆FgGET2 mutants were used in each of the subsequent experiments.

To restore the FgGET2 function in ∆FgGET2, a 2.4-kb region containing the coding sequence and its promoter was amplified by PCR using the primer pair C-FgGET2-F/R and then cloned into plasmid JM45. The resulting construct (JM45-FgGET2) was inserted into ∆FgGET2 protoplasts via polyethylene glycol-mediated transformation (Hou et al. 2004). G418 Sulfate (Calbiochem) was added for a final concentration of 300 mg/mL to screen for transformants. The generated C-FgGET2 transformants were identified by RT-PCR, with the primer pairs HPH-F/R, NEO-F/R, and M-FgGET2-F/R used to analyze hygromycin resistance gene, neomycin resistance gene, and FgGET2 expression, respectively.

To determine the subcellular localization of FgGET2, its coding sequence was amplified by PCR using the primer pair FgGET2-eGFP-F/R and then inserted into the pRFHUE-eGFP vector (Crespo-Sempere et al. 2011). The resulting recombinant vector was inserted into the WT strain via A. tumefaciens-mediated transformation as described above. The generated overexpression transformants were identified by PCR and sequencing.

Sequence analysis

The full GET2 gene sequences from different organisms and the encoded aa sequences were obtained from the National Center for Biotechnology Information database (https://www.ncbi.nlm.nih.gov/). All aa sequences were aligned using the CLUSTALW program (https://www.genome.jp/tools-bin/clustalw). Transmembrane helices in FgGET2 were predicted using TMHMM (http://servises.healthtech.dtu.dk/servise.php?TMHMM-2.0).

Yeast complementation assay

Yeast strain BY4741 and its ∆ScGET2 mutant were ordered from HORIZON (Dublin, Ireland). The full ScGET2 and FgGET2 coding sequences were amplified by PCR using BY4741 and F. graminearum and primer pairs pYES2-ScGET2-F/R and pYES2-FgGET2-F/R, respectively. The amplified products were ligated into the pYES2 vector (Thermo Fisher Scientific, Waltham, MA, USA). The resulting recombinant vectors (pYES2-ScGET2 and pYES2-FgGET2) were inserted into ∆ScGET2. The BY4741 and ∆ScGET2 strains transformed with the empty pYES2 vector were used as controls. Yeast transformants were selected on a synthetic medium lacking uracil (Clontech, Mountain View, CA, USA). For the complementation assay, tenfold serial dilutions of yeast transformants were added to solid YEPD medium supplemented with different chemicals, including hydroxyurea, tunicamycin, hygromycin, and CuSO₄ (concentrations indicated in figure legends). For the heat sensitivity assay, yeast strains were grown at 39°C. All experiments were repeated independently three times.

Fungal pathogenicity in wheat spikes and DON production

To evaluate FHB symptoms in wheat, susceptible Triticum aestivum cultivar ‘SM482’ plants were inoculated. Plants were grown in a climate-controlled greenhouse under a 12-h (25°C) day:12-h (20°C) night cycle. At the mid-anthesis stage, two fully developed florets of a single central spikelet were inoculated with 1 × 10³ conidia using a micropipette. The inoculated heads were capped with a plastic bag for 48 h to maintain moisture and incubated at 25°C. The fungal biomass in the inoculated spikelets at 2 dpi was measured by performing a qRT-PCR analysis. Visual FHB disease symptoms on 10–20 inoculated spikes per treatment were assessed at 12 dpi. To investigate pathogenicity changes in detail, fungal penetration behavior on cellophane membranes was examined as previously described (Lopez-Berges et al. 2010).

To determine the effect of FgGET2 on the accumulation of DON in wheat spikes, each floret of a whole spike was inoculated with 1 × 10³ conidia of WT or mutant strains. At least five heads per strain were analyzed. The infected spikes were harvested at 8 dpi and ground to a fine powder in liquid nitrogen. DON production was measured using a competitive ELISA-based DON detection kit (Mlbio, Shanghai, China) and a Multiskan Spectrum instrument (Thermo Fisher) as described by Qi et al. (2019). A two-stage protocol (Miller et al. 1986; Qi et al. 2012) was used to determine whether FgGET2 is required for the production of DON in liquid medium.

Microscopic examination

Hyphal morphology was examined using fresh WT and ∆FgGET2 mycelia collected from 2-day-old colonies on PDA medium. Mycelia were fixed in 2.5% (v/v) glutaraldehyde, dehydrated using a graded ethanol series, and embedded in Epon812 resin (Sigma-Aldrich, St. Louis, MO, USA). Ultrathin sections were prepared using an EM UC7 ultramicrotome (Leica, Wetzlar, Germany), after which specimens were stained with uranyl acetate and lead citrate. Hyphal morphology was observed using an HT7700 120 kV transmission electron microscope (Hitachi, Tokyo, Japan).

To examine vacuole morphology, the conidia of each strain were cultured in potato dextrose broth medium (Aobox) for 24 h. Hyphae were stained with 10 µM (final concentration) CMAC for 30 min at 37°C as previously described (Shoji et al. 2006). The F. graminearum strain expressing the FgGET2-eGFP fusion was utilized for localizing FgGET2 within cells. Conidia were collected from CMC medium and germinated in YEPD medium for 6 and 24 h, after which fluorescence was examined. ER-Tracker Red was used to label the ER of hyphae as described by Yun et al. (2020). Fluorescence was detected under UV light using an Olympus-BX63 fluorescence microscope with a cooled CCD camera (DP80; Olympus, Tokyo, Japan).

Statistical analysis

All experimental data were collected from three independent samples. The significance of the differences between treatments was determined according to Student’s t-test implemented using the Data Procession System program (version 12.01; Zhejiang University, Hangzhou, China).