The Endoplasmic Reticulum Cargo Receptor FgErv14 Regulates DON Production, Growth and Virulence in Fusarium graminearum

Fusarium graminearum is a plant filamentous pathogenic fungi and the predominant causal agent of Fusarium head blight (FHB) in cereals worldwide. The regulators of the secretory pathway contribute significantly to fungal mycotoxin synthesis, development, and virulence. However, their roles in these processes in F. graminearum remain poorly understood. Here, we identified and functionally characterized the endoplasmic reticulum (ER) cargo receptor FgErv14 in F. graminearum. Firstly, it was observed that FgErv14 is mainly localized in the ER. Then, we constructed the FgErv14 deletion mutant (ΔFgerv14) and found that the absence of the FgErv14 caused a serious reduction in vegetative growth, significant defects in asexual and sexual reproduction, and severely impaired virulence. Furthermore, we found that the ΔFgerv14 mutant exhibited a reduced expression of TRI genes and defective toxisome generation, both of which are critical for deoxynivalenol (DON) biosynthesis. Importantly, we found the green fluorescent protein (GFP)-tagged FgRud3 was dispersed in the cytoplasm, whereas GFP-FgSnc1-PEM was partially trapped in the late Golgi in ΔFgerv14 mutant. These results demonstrate that FgErv14 mediates anterograde ER-to-Golgi transport as well as late secretory Golgi-to-Plasma membrane transport and is necessary for DON biosynthesis, asexual and sexual reproduction, vegetative growth, and pathogenicity in F. graminearum.


Introduction
Fusarium graminearum (teleomorph: Gibberella zeae) is a devastating fungal pathogen that causes Fusarium head blight (FHB) in cereal crops, including wheat, barley, and maize [1][2][3]. FHB has been spreading worldwide, leading to severe yield losses [4,5]. In addition, F. graminearum produces a kind of trichothecene mycotoxin named deoxynivalenol (DON), which is toxic to humans and livestock by binding to the ribosome to inhibit the protein synthesis of eukaryotic organisms [6,7]. However, the effective control of FHB remains challenging, with management relying mainly on fungicides, which have many unwanted effects [8,9]. Therefore, understanding the underlying mechanisms responsible for the virulence in F. graminearum may facilitate the development of novel strategies and effective means for FHB control.
F. graminearum infects flowering wheat spikelets and grows through the rachis node at the bottom of each spikelet so as to colonize the entire wheat head [10]. During plant infection, F. graminearum produces DON, a secondary metabolite that acts as an important virulence factor to facilitate the expansion of the pathogen in wheat heads [11]. Moreover,

Fungal Strains and Growth Conditions
The F. graminearum gene single deletion mutants were generated from the parental wild-type (WT) PH-1 strain, and all fungal strains used in this study are listed in Supplementary Table S1. For FgERV14 disruptions, the endogenous copy of FgERV14 was replaced with a hygromycin resistance (Hyg R) cassette in WT strain PH-1. PCR and Southern blotting were used to identify ∆Fgerv14 deletion mutants [48]. To generate complemented Life 2022, 12, 799 3 of 16 strains, 927-bp FgERV14 5 promoter region and open reading frames were amplified by PCR from the genome of F. graminearum, then, the 3 end of the PCR product was fused with green fluorescent protein (GFP) and inserted into the pYF11 vector to generate a plasmid expressing FgErv14-GFP. The resulting pYF11-FgErv14-GFP plasmid was transferred into the ∆Fgerv14 mutant protoplasts using PEG-mediated fungal transformation. The transformants were selected on a G418-resistance medium and identified by PCR and GFP signals [49]. Strains expressing FgRud3-GFP/GFP-FgSnc1-PEM/FgErv14-GFP and FgSed5-mCherry/FgKex2-tdTomato/Tri4-mCherry under their respective native promoters were also obtained by PEG-mediated transformation and selected on a G418-resistance medium and nourseothricin (NAT)-resistance medium, respectively. V8 medium, 5 × yeast extract-glucose (YEG), complete medium (CM), or minimal medium (MM) were used for vegetative growth of fungal strains at 25 • C for 3 days [48]. The conidiation production was determined by culturing each strain in carboxymethyl cellulose (CMC) liquid medium. Perithecium and ascospore generation were induced by carrot agar medium. Trichothecene biosynthesis induction (TBI) medium was used to analyze trichothecene production. To examine mycelial growth under different stress conditions, CM supplemented with 1 M NaCl, 1 M sorbitol, 0.05% sodium dodecyl sulfate (SDS), and 5 mM H 2 O 2 were used separately [50]. Genomic DNA and RNA were extracted from mycelia harvested from all strains that were cultured in CM for 2 days.

Assessment of Conidia Germination and Perithecial Development
For conidia germination assay, five mycelial plugs (5 mm in diameter) of all strains were taken from the edge using a 50-mm punch and induced in CMC liquid medium at 25 • C for 5 days [51]. Then, the number of conidia was determined by the hemocytometer. The conidia germination was determined by suspending fresh conidia in yeast extract peptone dextrose (YEPD) liquid medium for 4 h, 6 h, and 8 h, and then observing them using the upright microscope (Eclipse Ni-U; Nikon, Tokyo, Japan) [48]. Sexual reproduction was assayed by first inoculating each strain on CA medium at 25 • C for 7 days. After removing the aerial hyphae, each strain was pressed down gently using 1 mL of 2.5% sterile Tween 20 solution [52]. The perithecium formation was determined and photographed after further inducing at 25 • C under black light for about 2-3 weeks.

Plant Infection Assays
Jimai 22 was used in this assay as it was a native susceptible wheat cultivar. Infection of flowering wheat inflorescences was performed as described previously [53]. Three-dayold conidia were collected and spore suspension with a concentration of 4 × 10 5 spores/mL was prepared. 10 µL of the suspension was injected into the flowering wheat spikelets. To maintain humidity of the wheat heads after injection, the ziplock bags were used after spraying with sterile distilled water. The bags were removed after 2 days, and the wheat heads showing symptoms of wheat head blight disease were collected and photographed two weeks post-inoculation. The disease index was quantified from 40 wheat heads of each strain incubation. For wheat coleoptile infection assay, 2.5 µL of three-day-old conidia suspensions (2 × 10 6 spores/mL) was used and the samples were photographed 8 days post-inoculation. The lesion length was quantified from 15 plants of each strain incubation.

DON Production and Quantitative Real-Time PCR (qRT-PCR) Analysis
DON biosynthesis was tested by culturing 3 fresh mycelial plugs of each strain on 5 g of wheat grains [54]. The production of DON and ergosterol production were analyzed using the plugs incubated at 25 • C for 20 days [50].
Total RNA extraction was performed as previously described [50]. The TRI genes expression levels were detected using qRT-PCR with the appropriate primers and liquid trichothecene biosynthesis induction (TBI) medium induction. The FgRPS16 gene in F. graminearum served as an internal control for normalization of gene expression levels [48], and data of qRT-PCR were calculated using the 2 −∆∆CT method [55].

Microscopic Observations
Each fungal strain expressing fluorescence-tagged proteins from plasmids and conidia stained with calcofluor white (CFW) were examined via fluorescence microscopy using confocal microscopy (LSM 880 NLO system; Zeiss, Oberkochen, Germany) or an upright fluorescence microscope (Eclipse Ni-U; Nikon) [27]. At least 10 different microscopic fields were observed for each sample with one or two channels (Zeiss confocal microscopy: 488 nm for GFP and 561 nm for mCherry/tdTomato. Nikon fluorescence microscopy: UV filter for CFW staining and GFP/RFP filters for fluorescence-tagged proteins). Images were acquired and analyzed using ImageJ and Adobe Photoshop.

Statistical Analyses
The experimental data are displayed as the mean ± standard deviation (SD) from at least three replicates. Duncan's test and analysis of variance (ANOVA) were performed using GraphPad Prism 8.0 to compare individual samples across different strains or treatments.

Identification and Knockout of the FgErv14 Protein in F. graminearum
The conserved integral membrane protein FgERV14 (FGSG_01398) of F. graminearum was detected using S. cerevisiae Erv14 (YGL054C) for a homology search. The FgERV14 gene is 629 bp in length and encodes a protein of 139 amino acids. From a protein alignment analysis, it was found that FgErv14 exhibited a high degree of identity to other fungal Erv14 proteins of Fusarium oxysporum (99.28%), Magnaporthe oryzae (92.70%), Neurospora crassa (91.97%), Aspergillus nidulans (87.59%), S. cerevisiae (62.69%) and Schizosaccharomyces pombe (45.54%) ( Figure S1A). A phylogenetic analysis showed that Erv14 is widespread in different eukaryotes ( Figure 1). Furthermore, a functional domain analysis revealed that Erv14 has three conserved transmembrane domains in fungi, animals, and plants ( Figure S1B). The amino acid sequence and domain similarity are more closely related to the FgErv14 protein, which might exhibit a conserved function in F. graminearum. To investigate the function of FgErv14, ∆Fgerv14 mutants were constructed using a homologous recombination strategy ( Figure S2A) and were identified by Southern blot analysis ( Figure S2B).

Involvement of FgErv14 in Vegetative Growth and Stress Resistance in F. graminearum
In S. cerevisiae, the absence of Erv14, a non-essential protein, does not affect vegetative growth [38,56]. The role of F. graminearum FgErv14 in vegetative growth was investigated by comparing WT PH-1, ∆Fgerv14 mutants (#3 and #4), and a ∆Fgerv14/FgERV14 complemented strain cultured on V8 juice medium. The ∆Fgerv14 mutant strains exhibited a phenotype of defective vegetative growth compared to that in PH-1 and the ∆Fgerv14/FgERV14 strain (Figure 2A,B). To further characterize whether the vegetative growth defects were medium dependent, each strain was grown on YEG, CM, and MM plates. Under all three conditions, the vegetative growth in ∆Fgerv14 mutants was also defective compared to that in PH-1 and the ∆Fgerv14/FgERV14 strain, consistent with the growth on the V8 medium. An examination of the responses of the mutants to osmotic stress (NaCl, sorbitol), cell wall disruption (SDS), and oxidative stress (H 2 O 2 ) revealed a significant inhibition of growth in the ∆Fgerv14 strain compared to WT and ∆Fgerv14/FgERV14 strains (Figure 3), suggesting that FgErv14 is important for osmoregulation and stress tolerance. To extend our detection, the hyphae of each strain were examined, and the ∆Fgerv14 mutant showed extensive branching hyphae in contrast to the WT and complementary strains ( Figure 2C). These results imply that FgErv14 regulates mycelial growth positively in F. graminearum.

Involvement of FgErv14 in Vegetative Growth and Stress Resistance in F. graminearum
In S. cerevisiae, the absence of Erv14, a non-essential protein, does not affect vegetative growth [38,56]. The role of F. graminearum FgErv14 in vegetative growth was investigated by comparing WT PH-1, ΔFgerv14 mutants (#3 and #4), and a ΔFgerv14/FgERV14 complemented strain cultured on V8 juice medium. The ΔFgerv14 mutant strains exhibited a phenotype of defective vegetative growth compared to that in PH-1 and the ΔFgerv14/FgERV14 strain (Figure 2A,B). To further characterize whether the vegetative growth defects were medium dependent, each strain was grown on YEG, CM, and MM plates. Under all three conditions, the vegetative growth in ΔFgerv14 mutants was also defective compared to that in PH-1 and the ΔFgerv14/FgERV14 strain, consistent with the growth on the V8 medium. An examination of the responses of the mutants to osmotic stress (NaCl, sorbitol), cell wall disruption (SDS), and oxidative stress (H2O2) revealed a significant inhibition of growth in the ΔFgerv14 strain compared to WT and ΔFgerv14/FgERV14 strains (Figure 3), suggesting that FgErv14 is important for osmoregulation and stress tolerance. To extend our detection, the hyphae of each strain were examined, and the ΔFgerv14 mutant showed extensive branching hyphae in contrast to the WT and complementary strains ( Figure 2C). These results imply that FgErv14 regulates mycelial growth positively in F. graminearum.

FgErv14 Controls F. graminearum Asexual and Sexual Reproduction
The role of FgErv14 in asexual development was explored by examining whether FgErv14 is required for conidial production. As shown in Figure 4A, after 5 days of culture on CMC medium, the conidial production of ∆Fgerv14 mutant was significantly lower than that of PH-1 and the ∆Fgerv14/FgERV14 strains. After 20 days, the conidial production of ∆Fgerv14 mutant was still much lower than that of PH-1 counted on the 5th day ( Figure 4A), indicating that FgErv14 plays an essential role in conidial production. To further investigate conidial morphology, CFW staining was used to observe the conidia of each strain. The data showed that 39.9% and 58.4% of conidia in the ∆fgerv14 mutant had 3 and 1-2 septa, respectively ( Figure 4B). By contrast, the conidia production in WT PH-1 and the ∆Fgerv14/FgERV14 strains were largely normal, with 4-7 septa produced by 85.9% and 85.2% of the conidia, respectively ( Figure 4B). Upon further analysis, the ∆Fgerv14 mutant was shown to exhibit delayed germ tube development compared to WT and ∆Fgerv14/FgERV14 strains ( Figure 4C), suggesting that FgErv14 is involved in germ tube germination and hence, in the conidial production in F. graminearum. each strain. The data showed that 39.9% and 58.4% of conidia in the Δfgerv14 mutant had 3 and 1-2 septa, respectively ( Figure 4B). By contrast, the conidia production in WT PH-1 and the ΔFgerv14/FgERV14 strains were largely normal, with 4-7 septa produced by 85.9% and 85.2% of the conidia, respectively ( Figure 4B). Upon further analysis, the ΔFgerv14 mutant was shown to exhibit delayed germ tube development compared to WT and ΔFgerv14/FgERV14 strains ( Figure 4C), suggesting that FgErv14 is involved in germ tube germination and hence, in the conidial production in F. graminearum.  The sexual reproduction of F. graminearum is the main step in the FHB disease cycle. Therefore, we examined perithecium generation in the ∆Fgerv14 mutant. PH-1 and ∆Fgerv14/FgERV14 strain growth on carrot agar medium for the sexual reproduction assay produced abundant perithecia at 2 weeks post-fertilization, whereas no perithecia were produced by the ∆Fgerv14 mutant ( Figure 4D), indicating that FgErv14 plays a crucial role in the sexual reproduction of F. graminearum.

FgErv14 Is Important for F. graminearum Virulence
To investigate the effects of FgErv14 in F. graminearum virulence, flowering wheat spikelets were inoculated by spore suspension with a concentration of 4 × 10 5 conidia/mL to test the virulence of each strain. At 2 weeks after inoculation, it was found that WT and the ∆Fgerv14/FgERV14 strain caused obvious symptoms of head blight, with an average disease index (the number of spikelets showing disease symptoms per head) of 12.94 and 11.89, respectively, indicating full virulence ( Figure 5A). By contrast, the ∆Fgerv14 mutant failed to spread around and no typical symptoms were observed in wheat flowering heads, with an average disease index of 0.05, indicating significantly attenuated virulence ( Figure 5A). Similar results were also obtained in the wheat coleoptiles assays. In contrast to the obvious lesion caused by PH-1 and the ∆Fgerv14/FgERV14 strain, the ∆Fgerv14 mutant induced limited lesion development only around the inoculation sites ( Figure 5B). These results re-establish that FgErv14 is involved in F. graminearum full virulence during plant infection.
ΔFgerv14/FgERV14 strain caused obvious symptoms of head blight, with an average disease index (the number of spikelets showing disease symptoms per head) of 12.94 and 11.89, respectively, indicating full virulence ( Figure 5A). By contrast, the ΔFgerv14 mutant failed to spread around and no typical symptoms were observed in wheat flowering heads, with an average disease index of 0.05, indicating significantly attenuated virulence ( Figure 5A). Similar results were also obtained in the wheat coleoptiles assays. In contrast to the obvious lesion caused by PH-1 and the ΔFgerv14/FgERV14 strain, the ΔFgerv14 mutant induced limited lesion development only around the inoculation sites ( Figure 5B). These results re-establish that FgErv14 is involved in F. graminearum full virulence during plant infection.

FgErv14 Is Necessary for the Production of DON in F. graminearum
A secondary metabolite mycotoxin DON is essential for F. graminearum virulence [11,57]. Therefore, we assessed DON biosynthesis in PH-1, the ΔFgerv14 mutant and

FgErv14 Is Necessary for the Production of DON in F. graminearum
A secondary metabolite mycotoxin DON is essential for F. graminearum virulence [11,57]. Therefore, we assessed DON biosynthesis in PH-1, the ∆Fgerv14 mutant and ∆Fgerv14/FgERV14 strains. As shown in Figure 6A, DON production by the ∆Fgerv14 mutant was 25% of that by WT PH-1 or the ∆Fgerv14/FgERV14 strain, with a DON level of only 0.71 mg ergosterol·mg −1 . The expression of some TRI genes responsible for the regulation of DON production were examined. Consistent with the reduction in the production of DON in the ∆Fgerv14 mutant, the qRT-PCR results showed that the expression levels of TRI1, TRI4, TRI5, TRI6, and TRI10 in the ∆Fgerv14 mutant were significantly lower than that of PH-1 ( Figure 6B), indicating that FgErv14 is essential for modulating the expression of TRI genes. Previous studies suggested that DON is synthesized in the toxisome localized in the ER [18,58]. Therefore, further monitoring of the toxisome formation was carried out by visualizing the localization of Tri4-GFP in each strain. As shown in Figure 6C, the Tri4-GFP signals from both PH-1 and ∆Fgerv14/FgERV14 strains were clustered in a crescent-shaped structure in hyphae, whereas no signal was observed in the ∆Fgerv14 mutant. Together, our results indicate that the defection of DON production in ∆Fgerv14 mutant was due to the essential role of FgErv14 in the formation of toxisome in F. graminearum.
genes. Previous studies suggested that DON is synthesized in the toxisome localized in the ER [18,58]. Therefore, further monitoring of the toxisome formation was carried out by visualizing the localization of Tri4-GFP in each strain. As shown in Figure. 6C, the Tri4-GFP signals from both PH-1 and ΔFgerv14/FgERV14 strains were clustered in a crescent-shaped structure in hyphae, whereas no signal was observed in the ΔFgerv14 mutant. Together, our results indicate that the defection of DON production in ΔFgerv14 mutant was due to the essential role of FgErv14 in the formation of toxisome in F. graminearum.

FgErv14 Is Mainly Localized in the ER and Toxisome in F. graminearum
The function of FgErv14 was further investigated by constructing the FgErv14-GFP fusion protein and then examining its cellular localization in F. graminearum. It has been reported that the S. cerevisiae Erv14 is mainly present on the ER [56]. Accordingly, a PH-1 strain co-expressed FgErv14-GFP with the ER marker FgSec61-tdTomato was obtained. The results showed that on CM, FgErv14-GFP colocalized with FgSec61-tdTomato ( Figure 7A), indicating that the subcellular localization of FgErv14 is in the ER, consistent with Erv14 in S. cerevisiae. Furthermore, as FgErv14 is involved in toxisome formation ( Figure 6C), strains that co-expressed FgErv14-GFP with toxisome marker Tri4-mCherry were constructed. Under toxin-inducing conditions, we found that FgErv14-GFP either accumulated in a crescent-shaped structure or dispersed in a circle within hyphae ( Figure 7B), and the FgErv14-GFP signals colocalized with Tri4-mCherry ( Figure 7B), indicating the presence of FgErv14 in the toxisome. These results indicate that toxin-inducing conditions is a key factor in the localization of FgErv14 to toxisomes. strain co-expressed FgErv14-GFP with the ER marker FgSec61-tdTomato was obtained. The results showed that on CM, FgErv14-GFP colocalized with FgSec61-tdTomato ( Figure  7A), indicating that the subcellular localization of FgErv14 is in the ER, consistent with Erv14 in S. cerevisiae. Furthermore, as FgErv14 is involved in toxisome formation ( Figure  6C), strains that co-expressed FgErv14-GFP with toxisome marker Tri4-mCherry were constructed. Under toxin-inducing conditions, we found that FgErv14-GFP either accumulated in a crescent-shaped structure or dispersed in a circle within hyphae ( Figure 7B), and the FgErv14-GFP signals colocalized with Tri4-mCherry ( Figure 7B), indicating the presence of FgErv14 in the toxisome. These results indicate that toxin-inducing conditions is a key factor in the localization of FgErv14 to toxisomes.

FgErv14 Is Required for FgRud3 and FgSnc1-PEM Localization in F. graminearum
In yeast, Erv14 is a component of COPII vesicles that regulate ER-to-Golgi transport [38]. During this process, Erv14, as an ER cargo receptor, regulates cargo protein loading during ER exit [56]. FgRud3 is an early-Golgi localization protein required for toxisome generation and virulence in F. graminearum [27]. Whether FgErv14 affects the localization of FgRud3 was determined by monitoring FgRud3-GFP signals in PH-1 and the ∆Fgerv14 mutant strains. In the former, FgRud3-GFP colocalized with mCherry-FgSed5 ( Figure 8A), the early-Golgi marker, whereas in the mutant lacking FgERV14, it was instead diffusely distributed throughout the cytosol ( Figure 8A), implying that the normal localization of FgRud3 depends on FgErv14. We then used late secretory pathway markers (GFP-FgSnc1-PEM: mutant V42A and M45A) to investigate whether FgErv14 is involved in the late secretory pathway, which mediates trafficking from late Golgi to PM [27]. In WT PH-1 hyphae, GFP-FgSnc1-PEM accumulated on the PM, whereas in hyphae of the ∆Fgerv14 mutant, although the signal was mostly localized to the PM, multiple green dots were also detected in the cytoplasm ( Figure 8B) colocalized with the late Golgi marker FgKex2-tdTomato, indicating that GFP-FgSnc1-PEM was partly trapped in the late Golgi.
These results suggest that there are defects in the Golgi-to-PM transport, but it was not completely blocked in the ∆Fgerv14 mutant.
distributed throughout the cytosol ( Figure 8A), implying that the normal localization of FgRud3 depends on FgErv14. We then used late secretory pathway markers (GFP-FgSnc1-PEM: mutant V42A and M45A) to investigate whether FgErv14 is involved in the late secretory pathway, which mediates trafficking from late Golgi to PM [27]. In WT PH-1 hyphae, GFP-FgSnc1-PEM accumulated on the PM, whereas in hyphae of the ΔFgerv14 mutant, although the signal was mostly localized to the PM, multiple green dots were also detected in the cytoplasm ( Figure 8B) colocalized with the late Golgi marker FgKex2-tdTomato, indicating that GFP-FgSnc1-PEM was partly trapped in the late Golgi. These results suggest that there are defects in the Golgi-to-PM transport, but it was not completely blocked in the ΔFgerv14 mutant.

Discussion
Vesicular transport from the ER to the Golgi in the early secretory pathway is necessary for growth and development in eukaryotes, including phytopathogenic fungi [27,34,36,37,59]. Erv14, an ER cargo receptor, circulates between the ER and Golgi apparatus and participates in the direct selection of COPII cargo proteins [38,43,[45][46][47]. Although Erv14 and its related proteins have been described in yeast and mammalian cells, their function in phytopathogenic fungi is still unclear. In this study, we identified the corresponding ER cargo receptor FgErv14 in F. graminearum. Our results demonstrate a role of FgErv14 in vesicle transport and thus, involvement in the control of the DON production, vegetative growth, asexual and sexual reproduction, and pathogenesis of F. graminearum.
In S. cerevisiae, ERV14 is a non-essential gene, and its absence does not affect vegetative growth [38,56]. However, in contrast to the phenotype of the erv14∆ yeast mutant, our results showed that the vegetative growth was reduced significantly in the ∆Fgerv14 mutant compared to WT (Figure 2A,B). Secretory trafficking is very important for the vegetative growth of filamentous fungi, and it was reported that a blockade of early secretory transport causes defects in vegetative growth in A. nidulans [60,61]. Therefore, it suggests that the growth defect of ∆Fgerv14 mutant is related to FgErv14-mediated secretory transport.
Moreover, the deletion of FgErv14 mutant was more hypersensitive to the osmotic, cell wall, and oxidative stresses (Figure 3), which is consistent with Erv14 deletion mutant results in greater sensitivity to osmotic stress in S. cerevisiae [46,47]. Thus, like S. cerevisiae Erv14, FgErv14 may affect partial Na + , K + /H + exchanger/antiporter (Nha1 [56]) localization by mediating vesicular trafficking, which leads to a more sensitive phenotype to osmotic and oxidative in ∆Fgerv14 mutant. As expected, the ∆Fgerv14 mutant produced significantly fewer conidia than WT, and the mutant also produced a higher proportion of morphological and germination defective conidia ( Figure 4B,C). Although this phenotype is consistent with those resulting from the partial absence of other COPII components and regulators, such as FgSec22 and FgRab1 [34,36], the defect level in the ∆Fgerv14 mutant was higher.
During the FHB disease cycle, the sexual reproduction of F. graminearum is the main source of infection of plants in nature [62,63]. Previous studies have shown that the COPII components FgSec22 and FgRab1 are not necessary for sexual reproduction in F. graminearum [34,36]. By contrast, we found that FgErv14 regulates perithecium formation and is required for sexual reproduction ( Figure 4D), suggesting that FgErv14 plays an important role in F. graminearum infection of plants. However, as the roles of FgSec22 and FgRab1 in regulating ER-to-Golgi transport in F. graminearum have not been determined, their contributions to early secretory transport during sexual reproduction is unclear. In yeast, Erv14 regulates prospore membrane formation and affects sexual reproduction through an as-yet unknown mechanism. The mechanism of FgErv14 regulating F. graminearum sexual reproduction needs to be further studied.
DON is the most toxic secondary metabolite produced by F. graminearum, and its production is affected by various physiological processes [16,64]. The protein encoded by the TRI genes are distributed in different steps to catalyze DON production [16]. During DON biosynthesis, its derivatives will also be produced [65]. In DON quantification, LC-MS technology is used to distinguish different derivatives. In the ∆Fgerv14 mutant, the expression of TRI gene was decreased, and the DON production was also greatly reduced in LC-MS determination. This result is consistent with many vesicle trafficking regulators, like Rab GTPases [30], FgVam7 [25], and FgLsb6 [66]. Together, FgErv14 plays a crucial role in DON biosynthesis.
The ER cargo receptor Erv14 is highly conserved in eukaryotes, but it does not mediate the transport of all newly synthesized proteins. The golgin protein Rud3, the general amino acid permease Gap1, and the Na + /H + antiporter are proteins transported by yeast Erv14 [45,56,67]. In this study, our results showed that FgErv14 was responsible for the localization of FgRud3 in the early-Golgi of F. graminearum ( Figure 8A), demonstrating evidence of its function as an ER cargo receptor that mediates the COPII vesicle-associated trafficking from the ER to Golgi. Our previous studies showed that FgRud3 regulates the pathogenicity of F. graminearum and affects toxisome formation in the mycelium, consistent with the phenotype of the ∆Fgerv14 mutant. Thus, at least some of the COPII cargo proteins selected by FgErv14 likely regulate pathogenicity and toxisome generation in F. graminearum, pointing to the regulatory role of transport of COPII cargo proteins. Under toxin-inducing conditions, FgErv14 is partially localized in the toxisome, suggesting its involvement in toxisome generation. We also observed that GFP-FgSnc1-PEM was partially trapped in the late Golgi apparatus in the ∆Fgerv14 mutant, whereas most of the GFP-FgSnc1-PEM is normally transported to the PM ( Figure 8B). According to these results, FgErv14 also affects late secretory transport from the late Golgi to the PM, but its absence does not completely block this pathway. In A. nidulans, the absence of SarA(Sar1) or RabO(Rab1) blocks ER-to-Golgi transport completely, resulting in the reabsorption of early Golgi cisternae into the ER, whereas late Golgi-associated proteins were dispersed in the cytoplasm [60,61,68]. Our results show that FgSed5 formed punctate structures and did not exhibit an ER localization phenotype in the ∆Fgerv14 mutant, indicating that FgErv14 does not regulate the localization of FgSed5. This result is consistent with the fact that FgErv14 does not mediate all COPII anterograde transport. Given that FgErv14 has only an apparent ER localization ( Figure 7A), we speculate that partial regulators of late secretory transport are abnormally localized in the ∆Fgerv14 mutant, resulting in defective late secretory transport. As secretory transport contributes significantly to the DON production, virulence, growth, and development of F. graminearum [28,69], FgErv14 may affect pathogenicity by regulating this process. Thus, these results support that FgErv14 regulates the transport of ER-to-Golgi and that this process affects Golgi-to-PM transport in F. graminearum.

Conclusions
In summary, our study demonstrated the involvement of the ER cargo receptor FgErv14 in vegetative growth, DON production, asexual and sexual reproduction, and pathogenicity in F. graminearum. FgErv14 was observed in the toxisome, where it plays an important role in toxisome generation. We also found that FgErv14 participates in ER-to-Golgi transport and thus, in the regulation of F. graminearum virulence.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/life12060799/s1, reference [70] is cited in the Supplementary Materials. Figure S1: Amino-acid sequence and domain alignment of Erv14; Figure S2: Generation and verification of FgERV14-deletion mutants; Table S1: F. graminearum strains used in this study.