Identifying transcription factors associated with Fusarium graminearum virus 2 accumulation in Fusarium graminearum by phenome-based investigation

Highlights

• •FgV2 was transferred to F. graminearum transcription factor deletion mutant library.
• •Defective RNAs adversely affected FgV2 accumulation and vertical transmission.
• •Some TFs were involved in the transcription of FgDICER-2 and FgAGO-1 against FgV2.
• •FgV2 infection depressed the resistance to hydroxyurea of TF mutant.
• •FgV2 accumulation was lower in ROS-excessive TF mutants compared to WT.

Abstract

Fusarium graminearum virus 2 (FgV2) infection induces phenotypic changes like reduction of growth rate and virulence with an alteration of the transcriptome, including various transcription factor (TFs) gene transcripts in Fusarium graminearum. Transcription factors are the primary regulator in many cellular processes and are significant in virus-host interactions. However, a detailed study about specific TFs to understand interactions between FgV2 and F. graminearum has yet to be conducted. We transferred FgV2 to a F. graminearum TF gene deletion mutant library to identify host TFs related to FgV2 infection. FgV2-infected TF mutants were classified into three groups depending on colony growth. The FgV2 accumulation level was generally higher in TF mutants showing more reduced growth. Among these FgV2-infected TF mutants, we found several possible TFs that might be involved in FgV2 accumulation, generation of defective interfering RNAs, and transcriptional regulation of FgDICER-2 and FgAGO-1 in response to virus infection. We also investigated the relation between FgV2 accumulation and production of reactive oxygen species (ROS) and DNA damage in fungal host cells by using DNA damage- or ROS-responsive TF deletion mutants. Our studies provide insights into the host factors related to FgV2 infection and bases for further investigation to understand interactions between FgV2 and F. graminearum.

1. Introduction

Fusarium graminearum causes enormous economic losses as a significant pathogen in wheat and maize (Savary et al., 2019). In addition, this filamentous fungus and related Fusarium species produce a variety of mycotoxins that are hazardous to human and animal health (Hof, 2020). For these reasons, researchers have investigated the mechanisms of pathogenicity and virulence to manage this pathogen. Loss-of-function studies analyzed important phenotypic traits such as growth, virulence, and biosynthesis of mycotoxins (Chen et al., 2019; Son et al., 2011; Yu et al., 2014). Although chemical fungicide has been mainly used for managing F. graminearum, RNAi-based strategies like spray-induced gene silencing and host-induced gene silencing, targeting essential genes for F. graminearum, were recently established for disease management (Wang et al., 2020; Werner et al., 2020).

Transcription factors (TFs) are regulatory proteins that control gene expression mainly by binding to DNA. These TFs construct the hierarchy and network, then finely manage many genes to respond to biotic and abiotic stresses (Song et al., 2016). Including F. graminearum, some studies have examined the regulatory function of TFs in cellular processes in plant pathogenic fungi (John et al., 2021). The gene deletion library of 657 putative TFs of F. graminearum was constructed, and their function was analyzed based on the phenotypes of each TF mutant (Son et al., 2011). They analyzed putative F. graminearum TFs involved in important phenotypic traits such as sexual development, mycotoxin production, and growth and identified their interconnection. Recently, the potential roles of TFs belonging to the basic leucine zipper (bZIP) family were demonstrated under various abiotic stresses and development stages (Hussain et al., 2022).

Fusarium graminearum virus 2 (FgV2), which belongs to the genus Betachrysovirus, has five segmented double-stranded RNAs (dsRNAs) as a genome (Kotta-Loizou et al., 2020; Yu et al., 2011). Each dsRNA of FgV2 encodes a single open reading frame (ORF) and shares a high similarity of nucleotide and amino acid sequences with Fusarium graminearum virus China 9 (FgV-ch9) (Darissa et al., 2011; Yu et al., 2011). ORF1 encodes a protein that contains a typical RNA-dependent RNA polymerase (RdRp) domain, and ORF3 is estimated to encode the capsid protein (CP) subunit. FgV-ch9 and FgV2 infection cause noticeable phenotypic changes like growth retardation, reduced virulence, and changes in gene expression of F. graminearum (Darissa et al., 2011; Lee et al., 2014). Previous research found that one of the putative mRNA binding proteins, virus response 1 (vr1, locus tag FGSG_05737), is involved in FgV-ch9-related symptom severity in F. graminearum (Bormann et al., 2018). Recently, it was demonstrated that P2 and P3 CPs of FgV-ch9 were expressed at full-length size and underwent proteolytic processing driven by unknown host factors (Lutz et al., 2021). However, host factors that might be related to FgV2 are mainly unknown.

Several researches have been conducted to understand the relationship between mycovirus and its subviral RNAs in the fungal host. It was reported that the host dicer and argonaut genes, responsible for RNA silencing against Cryphonectria hypovirus 1 (CHV1) infection, are required for CHV1 defective RNA production, promoting viral RNA recombination (Sun et al., 2009b; Zhang and Nuss, 2008). With Rosellinia necatrix partitivirus 2 and Rosellinia necatrix partitivirus 6, the effect of defective-interfering RNAs on its parental virus and the fungal host was assessed (Chiba et al., 2013, 2016). However, studies on defective RNAs in mycoviruses infecting Fusarium spp. have yet to be conducted.

Studies about associations between TFs and mycoviruses are essential to understand host-mycovirus interactions, but available information is still limited. Phenome-based analysis using F. graminearum TFs deletion mutants was conducted to identify TFs related to Fusarium graminearum virus 1 (FgV1) replication and virus-infected symptom development (Yu and Kim, 2021). FgV1 infection also causes hypovirulent effects such as a reduction of growth and virulence in F. graminearum (Lee et al., 2014). However, FgV1 has a single-stranded RNA genome and is phylogenetically distant from FgV2 (Cho et al., 2013; Honda et al., 2020; Kwon et al., 2007). In addition, the transcription pattern following FgV2 infection in F. graminearum was significantly different from FgV1 infection, suggesting that FgV2 might have distinct strategies for its survival (Lee et al., 2014). In this study, we transferred FgV2 to 657 putative TF gene deletion mutant library of F. graminearum to investigate the associations between specific TFs and FgV2 infection. Our data can provide broader insights into fungal host transcription factors' role(s) on mycovirus infection and clues to comprehending mycovirus-host relationships.

2. Materials and methods

2.1. Fungal strain and culture conditions

Wild-type (WT) and 657 TF gene deletion mutant library of F. graminearum GZ03639 strain were provided by the Center for Fungal Genetic Resources in South Korea. TF deletion mutants were activated from 25% glycerol stock, which was stored at −80 °C, on potato dextrose agar (PDA) containing 50 ppm geneticin. Then, grown TF mutant colonies were subcultured on complete medium (CM) agar with geneticin. Fungal mycelia were cultured in CM broth for 5 days at 25 °C shaking incubator with 150 rpm. Grown mycelia were harvested by filtering with 3 MM paper and then washed with distilled water. After removing water using a paper towel, mycelia were ground with liquid nitrogen and stored at −80 °C for further study.

2.2. Virus transmission

FgV2-infected F. graminearum GZ03639 (WT-VI) was co-cultured on CM agar with TF mutants to transfer FgV2 via hyphal anastomosis. Mycelia from the fused region where two fungal colonies encountered were subcultured 2–3 times on CM agar containing geneticin to get stable FgV2-infected TF mutant colonies. FgV2-infected colonies were selected based on phenotypic changes. When the virus was not transmitted to TF mutants through hyphal anastomosis despite several trials, virus transmission was attempted through protoplast fusion. Protoplast fusion was performed as described previously with modifications (Lee et al., 2011). Young mycelia of FgV2-infected WT and TF mutants were incubated with 1 M NH₄Cl containing 1% (w/v) driselase (Sigma-Aldrich, USA) at 30 °C. Protoplasts were harvested by centrifugation, washed with STC (1.2 M Sorbitol, 10 mM Tris–HCl pH 7.5, 50 mM CaCl₂), and suspended in 200 µl of MMC buffer (0.6 M Mannitol, 10 mM MOPS pH 7.0, 10 mM CaCl₂). Protoplast suspensions from FgV2-infected WT and TF mutants were mixed with the proportion of 2:1 and placed on ice for 20 min. Protoplast fusion was mediated by 500 µl of polyethylene glycol solution (60% PEG 3350, 10 mM MOPS pH 7.0, and 10 mM CaCl₂), and the mixture was incubated at 25 °C for 20 min. Fused protoplasts were regenerated in 700 µl of TB3 broth for 7 days at 25 °C and plated on CM agar containing 50 µg/ml geneticin. FgV2-infected antibiotic-resistant colonies were screened and subcultured twice. FgV2 infection was confirmed by detecting dsRNA obtained through treatment of DNase Ⅰ (Takara Bio, Japan) and S1 nuclease (Takara Bio, Japan) to 3 µg of total RNA on 1% agarose gel. When FgV2 was not transmitted by hyphal anastomosis or protoplast fusion despite three times trials, these TF mutants were determined as FgV2 non-transmitted.

2.3. Measurement of fungal isolates growth

The colony morphology of FgV2-infected TF deletion mutants was photographed after 5 days of incubation on CM agar (Supporting Information Fig. S1). The radial growth of each TF mutant to 4 directions was measured using the Image J program (Collins, 2007). The mean growth rate was calculated with at least 3 biological replicates of TF deletion mutants.

2.4. Total RNA extraction and cDNA synthesis

The preparation of total RNA samples and cDNA synthesis were performed as described previously (Yu et al., 2020). RNAisoplus (Takara Bio, Japan) was added to ground mycelia, and total RNA was extracted following the manufacturer's instructions. Extracted total RNA was treated by DNase Ⅰ (Takara Bio, Japan) to remove genomic DNA and precipitated by ethanol and NaOAc. After washing with 80% ethanol, the total RNA was resuspended in DEPC-water. Five micrograms of DNase Ⅰ-treated total RNA were used for cDNA synthesis with 50 pmol oligo(dT)₁₈ primers. Reverse transcription was performed by GoScript™ reverse transcriptase (Promega, USA), and synthesized cDNAs were diluted to 20 ng/μl with distilled water for RT-qPCR and semi-quantitative RT-PCR analysis.

2.5. Estimation of FgV2 dsRNA accumulation

To detect the FgV2 dsRNA accumulation level, 3 µg of DNase Ⅰ-treated total RNA was treated by S1 nuclease (Takara Bio, Japan). dsRNA fraction was precipitated by ethanol and resuspended in DEPC-water. For measuring FgV2 dsRNA accumulation, isolated dsRNA and 3 µg of DNase Ⅰ-treated total RNA were loaded on 1% agarose gel with FgV2-infected WT samples as a control and analyzed by gel electrophoresis with ethidium bromide staining. RNA band intensity was measured by the Image J program. FgV2 dsRNA amount relative to 18S rRNA was calculated and standardized by the value of WT-VI control. In the case of detecting defective interfering RNAs, these defective interfering RNAs were excluded when measuring the FgV2 dsRNA accumulation level.

2.6. RT-qPCR and semi-quantitative RT-PCR

RT-qPCR was performed using a Bio-Rad CFX384 real-time PCR system. PCR reaction mixture (20 μl) was prepared with 20 ng of cDNA, 10 μl of 2X Real-Time PCR Master Mix For SYBR® Green I (BioFACT, Korea), and 10 pmoles of each forward and reverse primer. The reaction condition was as follows: 15 min at 95 °C and 40 cycles of 20 s at 95 °C, 30 s at 57 °C, and 20 s at 72 °C, and the melting curve was generated by raising the temperature from 55 to 95 °C. Two reference genes, EF1α (FGSG_08811) and UBH (FGSG_01231), were used as internal controls (Kim and Yun, 2011; Son et al., 2013). The experiment was conducted with 3 biological replicates. The obtained data were analyzed using Bio-Rad CFX Manager 3.0 software. Semi-quantitative RT-PCR was performed with a 20 μl reaction mixture containing 20 ng of cDNA, TaKaRa Taq™ (Takara Bio, Japan), and 10 pmoles of each primer set. The reaction condition was as follows: 3 min at 95 °C and 25 or 30 cycles of 25 s at 95 °C, 30 s at 57 °C, and 20 s at 72 °C, and the final extension was conducted at 72 °C for 7 min. Amplified PCR product was visualized on 1% agarose gel containing ethidium bromide under UV light after gel electrophoresis. All primer sets are listed in Table S2.

2.7. Defective interfering RNA extraction and sequence analysis

Thirty-microgram of total RNA was treated with DNase Ⅰ (Takara Bio, Japan) and S1 nuclease (Takara Bio, Japan) and separated on 1% agarose gel by gel electrophoresis. Purifying defective interfering RNA from agarose gel was conducted by repeated cycles of freezing in liquid nitrogen and thawing in a 42 °C water bath. The aqueous phase was obtained with 1:1 (v:v) acid-phenol:chloroform (Ambion, USA). dsRNAs in the aqueous phase were recovered by ethanol precipitation and centrifugation. The obtained dsRNA pellet was suspended in DEPC-water. cDNA was synthesized using GoScript^TM reverse transcriptase (Promega, USA) with random hexamer or primer mixture, including 3′-UTR sequences of each FgV2 RNA. cDNA was amplified by PCR with a random or virus-specific primer set for the UTR of each FgV2 RNA. Primer sets are listed in Table S2. PCR product was cloned to pGEM-T easy vector (Promega, USA), and the sequence was analyzed by Macrogen (Seoul, South Korea). Sequences of defective interfering RNA were aligned with FgV2 RNAs using SeqMan software (DNASTAR, Madison, WI, USA)

2.8. Vertical transmission of FgV2 in the fungal host through asexual reproduction

Conidiation was induced by incubating FgV2-infected WT and TF mutant in carboxymethyl cellulose (CMC) broth for 5 days at 25 °C with 150 rpm shaking. Each conidial suspension was diluted approximately to 100 conidia/ml and spread on CM agar media. After incubation at 25 °C for at least 2 days, germinated single conidia were transferred to CM or geneticin-containing CM agar media for WT or TF mutants, respectively. Newly transferred colonies were incubated for 5 days at 25 °C and grouped according to colony size and morphology. FgV2 infection of each group was confirmed by dsRNA profile on 1% agarose gel after total RNA digestion with DNase Ⅰ (Takara Bio, Japan) and S1 nuclease (Takara Bio, Japan). After separating FgV2-infected colonies according to the presence of defective interfering RNAs, mycelial plugs from several FgV2-infected colonies were inoculated into CMC broth for vertical transmission to the next generation.

2.9. ROS detection

WT and TF mutants were incubated on CM agar for 3 days at 25 °C. To soak the colony, one milliliter of 0.2% nitrotetrazolium blue chloride (NBT) solution was added to each plate. After incubation for 30 min in dark conditions, the NBT solution was drained, and the colony was washed with ethanol. A stained colony picture was taken after more than 30 min incubation.

2.10. Protein-protein interaction (PPI) network of transcription factors associated with the FgV2 infection

TFs whose gene was differentially expressed upon FgV1 or FgV2 infection (log2 fold changes>0.5) were selected to establish the PPI network. PPI of TF proteins was analyzed by the Cytoscape STRING program (0.4 confidence cutoff) with RNA-Seq data of each TF gene (Doncheva et al., 2018; Lee et al., 2014). Among the PPI network, the largest network was selected. Functional annotation of TF proteins from the PPI network was conducted using FungiFun 2.2.8 with FunCat classification ontology (Priebe et al., 2015).

3. Results

3.1. Phenotypic analysis of FgV2-infected TF gene deletion mutant library

We transferred FgV2 to the 657 TF gene deletion mutant library to investigate the relationship between each TF gene and FgV2 infection in F. graminearum. Firstly, FgV2 transmission was conducted through anastomosis from FgV2-infected F. graminearum GZ03639 (WT-VI) to the virus-free TF gene deletion mutant library. Among a total of 657 TF mutants, FgV2 was transmitted into 639 TF mutants via hyphal anastomosis or protoplast fusion; however, 18 TF mutants were unable to obtain virus-infected strains despite repeated trials. For phenotypic analysis of FgV2-infected TF mutants, the colony morphology of FgV2-infected TF mutants was photographed. A representative colony morphology image of each FgV2-infected TF mutant was shown in Fig. S1.

Although most FgV2-infected TF mutants showed similar phenotypic changes with WT-VI, such as reduced radial growth and aerial mycelia in general, FgV2-infected TF mutants can be classified into 3 groups according to relative mycelial growth compared to virus-free wild-type (WT-VF) or virus-free TF mutants. When the average radial growth of the WT-VF colony as a mycelial length was set to 100, WT-VI showed 52–89 (average 68.5 ± 11.5) radial growth. The mycelial length of FgV2-infected TF mutants was calculated with the same method. In the case of TF mutants that showed reduced mycelial growth without virus infection, the mycelial length of these virus-free TF mutants was used as a standard for measurement of the relative radial growth of FgV2-infected TF mutants. We classified FgV2-infected TF mutants showing growth below 45 or over 90 upon FgV2 infection compared to virus-free TF mutants into Group 1 or Group 3, respectively. TF mutants showing 45–90 mycelial length on FgV2 infection, similar to the mycelial length of WT-VI against WT-VF, were classified into Group 2 (Fig. 1).

Fig. 1.

Representative colony morphology of FgV2-infected TF deletion mutants in each group. All colony pictures were taken after 5 days of incubation on CM agar media. WT means wild-type Fusarium graminearum strain GZ03639. The virus-free colony morphology of TF deletion mutants in this figure was almost identical to virus-free WT.

Most of the FgV2-infected TF mutants, approximately 85%, were classified into Group 2, showing a similar level of growth reduction to WT-VI, while few FgV2-infected TF mutants were in Group 1 and Group 3, approximately 8% and 7%, respectively. Relatively high phenotypic variation was observed in TF mutants belonging to bZIP, GATA type zinc finger, and Myb showing that more than 20% of FgV2-infected TF mutants in each TF family were classified into Group 1 or Group 3 (Table 1).

Table 1.

Dissection of FgV2-infected transcription factor deletion mutants.

TF Classification	TF	ΔTF	ΔTF-FgV2	Group
				1	2	3
bHLH	16	15	14a	2	12	0
bZIPb	22	22	22	3	17	2
C2H2 zinc finger	98	94	94	6	83	5
Heteromeric CCAAT	8	8	8	0	8	0
HMG	37	34	33 a	2	29	2
Homeodomain-like	14	7	7	0	7	0
Nucleic acid-binding, OB-fold	47	40	40	0	34	6
Winged helix repressor DNA-binding	27	26	25 a	0	23	2
Helix-turn-helix, AraC type	8	7	7	0	7	0
GATA-type zinc fingerb	8	7	7	1	5	1
Zinc finger, CCHC-type	12	12	12	0	12	0
Zn2Cys6 zinc finger	316	296	284 a	32	242	10
Mybb	19	17	17	1	12	4
Others	77	72	69 a	3	53	13
Total	709	657	639	50	544	45

^aFgV2 was not transmitted to some TF mutants through hyphal fusion or protoplast fusion.

^bRelatively high phenotypic variation was observed in some TF families showing that more than 20% of FgV2-infected TF mutants in each TF family were classified into Group 1 or Group 3.

Among 95 TF deletion mutants in Groups 1 or 3 in FgV2-infected strains, 16 TF deletion mutants showed similar mycelial growth patterns with FgV1-infected TF deletion mutants (Fig. S2). In contrast, some TF deletion mutants showed the opposite colony growth pattern followed by FgV1 or FgV2 infection; for example, ΔGzZC021, which belongs to Group 1 in FgV2-infected ΔTF, was classified into Group 3 in FgV1-infected ΔTF (Fig. S2). ΔGzOB007, ΔGzOB008, and ΔGzZC273 were in Group 3 in FgV2-infected ΔTF but in Group 1 in FgV1-infected ΔTF. We also examined the target gene expression levels in WT-VI to confirm that TF genes were affected by FgV2 infection (Table S3). Among these, the transcription level of the GzZC059 gene, which was classified into Group 1 in both FgV1 or FgV2 infection, was significantly increased by both FgV1- or FgV2-infected WT (Table S3, (Lee et al., 2014)). Our observations showed that single TF gene deletion has varied effects on fungal colony morphology depending on FgV1 or FgV2 infection.

3.2. Relationship between mycelial growth and FgV2 accumulation in TF mutants

Previously, it was reported that FgV1 ssRNA and dsRNA accumulation levels were inversely correlated with mycelial growth in FgV1-infected TF deletion mutants (Yu and Kim, 2021). Therefore, to investigate the relationship between mycelial growth and FgV2 accumulation level in TF mutants, we detected the dsRNA accumulation level in FgV2-infected TF deletion mutants. Then, the value of the dsRNA accumulation level was plotted according to the mycelial growth of each FgV2-infected TF mutant. As a result, we observed a tendency that the FgV2 dsRNA accumulation level was high as the mycelial growth rate was reduced (Fig. 2). In some FgV2-infected TF mutants, defective interfering RNAs (DI-RNAs) were detected together with viral genomic RNA. Many DI-RNAs were often generated and significantly interfered with viral genomic RNA accumulation in Groups 2 or 3. However, DI-RNA generation and the interfering effect of DI-RNAs were relatively subtle in TFs in Group 1 (Fig. S3). These observations suggest that some TFs might affect FgV2 viral RNA accumulation processes and virus-host interaction.

Fig. 2.

Relationship between mycelial growth and dsRNA accumulation level in FgV2-infected TF mutants. A total of 33 TF mutants with WT were plotted according to dsRNA accumulation level and mycelial length (10 for Group 1, 12 for Group 2, and 11 for Group 3). In the case of DI-RNA detection at least 1 replicate of 3, TF was indicated with an orange dot.

3.3. Defective interfering RNAs in FgV2-infected TF gene deletion mutants

In the process of confirming the FgV2 infection, defective interfering RNAs were detected together with FgV2 gRNAs in WT-VI and several TF mutants. Reduced viral gRNAs by the DI-RNA resulted in a faster growth rate of fungal colonies than DI-RNA-free colonies (Fig. 3A and B). In WT-VI, DI-RNAs produced from FgV2 RNA3 were detected, although they were removed during the subculture (Fig. 3C). Because only DI-RNA-harboring colonies were collected from some TF mutants in which DI-RNAs were generated, we selected one TF mutant (ΔGzZC138) in which DI-RNA-free and DI-RNA-harboring colonies were collected together in FgV2 infection to investigate the effect of DI-RNAs on the virus and fungal host. DI-RNAs in FgV2-infected ΔGzZC138 were about 1.5 kb generated from FgV2 RNA3, and FgV2 gRNA accumulation level was significantly reduced in the presence of DI-RNA like WT (Fig. 3C and D). To investigate the effect of DI-RNAs on FgV2 stability in the host, we conducted vertical transmission via asexual reproduction. In DI-RNA-free WT-VI, the virus transmission rate reached 100% in the 2nd generation without the production of DI-RNAs despite a low transmission rate in the 1st generation, showing consistent results with the previous report (Fig. 3E; Lee et al., 2014). In the DI-RNA-free FgV2-infected ΔGzZC138, the transmission rate was low compared to that of DI-RNA-free WT-VI, but it was increased following serial passage without the spontaneous production of DI-RNAs (Fig. 3E). However, in DI-RNA-harboring FgV2-infected ΔGzZC138, the FgV2 was not stably transmitted via asexual reproduction to the 3rd generation and DI-RNA-harboring colonies were generated constantly. With the DI-RNA, the virus transmission rate in the 3rd generation was significantly low compared to DI-RNA-free WT-VI or ΔGzZC138. The negative effect of DI-RNAs on virus transmission rate was also observed in other DI-RNA-harboring FgV2-infected TF mutants, ΔFgAT003 and ΔGzZC069 (Fig. S4). These results indicated that the deletion of some of TFs could affect the generation of DI-RNA and that DI-RNAs adversely affect FgV2 gRNA accumulation and vertical transmission in its host.

Fig. 3.

Defective RNAs in WT and FgV2-infected ΔGzZC138. (A-B) Colony morphology of FgV2-infected WT (A) and ΔGzZC138 (B) depending on DI-RNA presence. Colony picture was taken after 5 days inoculation on CM agar media. (C) Relative FgV2 accumulation in WT and ΔGzZC138 depending on DI-RNA presence. Three micrograms of total RNA were digested by DNase І and S1 nuclease and separated on 1% agarose gel (Left). Lane M is a 1 kb DNA marker (Bioneer, Korea). Relative FgV2 gRNA accumulation level was measured by Image J (Right). Asterisk represents FgV2 accumulation level is statistically different (P < 0.05) based on the LSD test. VI-DI means DI-RNA-harboring FgV2-infected. DI-RNA was marked with an arrow. (D) Schematic diagram of DI-RNAs generated from FgV2 RNA3 in FgV2-infected WT-VI and ΔGzZC138. The sequence confirmed region was indicated with red bars. (E) The vertical transmission rate of FgV2 via asexual reproduction in WT and ΔGzZC138 depending on DI-RNA presence. 1 G to 3 G, 1st generation to 3rd generation, respectively; VF, virus-free colony; VI/DI-, FgV2-infected colony without DI-RNA; VI/DI+, FgV2-infected colony harboring DI-RNA.

3.4. Association of TFs to transcription of FgDICER-2 and FgAGO-1

RNAi system is the most well-known antiviral mechanism in fungi (Lax et al., 2020). A previous study demonstrated that FgV2 infection evokes transcriptional induction of FgDICER-2 and FgAGO-1, which have a pivotal role in RNAi responsive pathway against mycovirus infection and hairpin-induction in F. graminearum (Yu et al., 2018). We hypothesized that there might be some TFs involved in the transcriptional induction of FgDICER-2 and FgAGO-1 in the RNAi-responsive pathway following FgV2 infection and that deletion of a transcription factor might affect the induction of two RNAi-related genes. To investigate the association of TFs in RNAi responses, we analyzed the transcriptional changes of FgDICER-2 and FgAGO-1 in several TF deletion mutants in Groups 1 and 3, in which FgV2 accumulation was previously confirmed in Fig. 2 (Table 2). As a result, we found that the relative gene expression of FgDICER-2 and FgAGO-1 was significantly induced in FgV2-infected TF mutants in Group 1 compared to WT-VF, on the contrary, it didn't show significant differences in the expression level of FgDICER-2 and FgAGO-1 in several FgV2-infected TF mutants in Group 3 compared to those of WT-VF (Table 2). Because the transcriptional induction of FgDICER-2 and FgAGO-1 tends to increase with the amount of viral dsRNA, differences in transcriptional induction level of FgDICER-2 and FgAGO-1 between Groups 1 and 3 are general aspects. Interestingly, the expression level of FgDICER-2 and FgAGO-1 in FgV2-infected ΔGzZC137, ΔGzGATA007, and ΔGzMIZ001 showed different expression patterns compared to WT-VI or other TF mutants in the same group. This result suggested the possible involvement of some transcription factors that regulate the transcription of FgDICER-2 and FgAGO-1 in response to virus infection.

Table 2.

Relative FgDICER-2 and FgAGO-1 gene expression levels in FgV2-infected TF mutants in Groups 1 and 3.

		FgDICER-2a	FgAGO-1a
Group 1	WT-VI	11.17±2.13*	5.74±1.49*
	ΔGzbZIP008-VI	22.42±11.18*	5.56±1.60*
	ΔGzHOME003-VI	10.45±4.58*	4.62±1.12*
	ΔGzZC009-VI	17.08±3.63*	6.04±1.11*
	ΔGzZC014-VI	22.68±6.54*	8.00±2.02*
	ΔGzZC039-VI	23.17±2.09*	4.59±0.85*
	ΔGzZC058-VI	29.69±15.48*	7.01±3.61*
	ΔGzZC059-VI	25.17±6.10*	6.09±1.63*
	ΔGzZC137-VI	2.93±1.56	1.48±0.65
	ΔGzZC176-VI	14.30±9.40*	4.02±1.57*
	ΔGzZC188-VI	26.38±12.32*	7.00±2.79*
Group 3	WT-VI	12.86±3.16*	6.58±1.25*
	ΔFgStuA-VI	0.63±0.45	2.13±0.87
	ΔGzGATA007-VI	12.46±4.31*	7.15±1.91*
	ΔGzMADS003-VI	2.22±0.94	1.56±0.29
	ΔGzMyb008-VI	3.13±1.26	2.35±1.17
	ΔGzMyb016-VI	0.97±0.74	0.75±0.24
	ΔGzMIZ001-VI	14.17±3.76*	3.45±0.89*

^aQuantification of relative gene expression level compared to WT-VF. cDNAs were synthesized from 5 micrograms of total RNA samples extracted after 5 days of incubation. Two reference genes, EF1α (FGSG_08811) and UBH (FGSG_01231), were used as internal controls. The experiment was conducted with three biological replicates.

^⁎Asterisk means the value significantly differed (P < 0.05) compared to the value of WT-VF based on the LSD test.

3.5. Association of TFs involved in DNA damage response or oxidative stress response with FgV2 infection

Phenotypes of TF gene deletion mutants were assessed under varied abiotic stress conditions, and several TF mutants sensitive to each stress condition were identified (Son et al., 2011). In this study, we compared the colony growth under various abiotic stress depending on FgV1 or FgV2 infection to explore the possible association of virus infection with abiotic stress response in F. graminearum (Fig. 4A). As a result, we found that colony growth was significantly hampered, followed by virus infection under some stress conditions, such as osmotic and DNA-damage conditions, especially with hydroxyurea (HU). In a previous study, we also observed that a relatively high number of DNA-damage response (DDR)-related TF deletion mutants showed significant changes in colony morphology following FgV1 infection (Yu and Kim, 2021). To investigate the possible association of FgV2 infection with DNA damage in F. graminearum, we selected several TF mutants showing different sensitivity to DNA-damaging chemicals and assessed colony growth and FgV2 accumulation level in HU-containing media. Interestingly, in HU-containing media, FgV2 infection caused a noticeable growth reduction in ΔGzGATA007, which is resistant to HU in a virus-free state compared to WT (Fig. 4B). In addition, the virus accumulation level was significantly increased in ΔGzGATA007 with HU (Fig. 4C). FgV2 accumulation level was decreased with HU in ΔGzscp and ΔRFX1, showing more sensitivity to HU. In ΔGzZC303, which is similarly sensitive to HU with WT, colony growth and FgV2 accumulation showed a similar pattern with WT.

Fig. 4.

(A) Colony morphology of FgV1 or FgV2-infected WT under various abiotic stress conditions. Gz03639, Gz03639/FgV1, and Gz03639/FgV2 mean WT-VF, FgV1-infected WT, and FgV2-infected WT, respectively. CR and SDS represent Congo red and sodium dodecyl sulfate, respectively. Colony pictures were taken 5 days after inoculation on each media. (B) Comparison of colony morphology of WT or DDR-related TF mutants under DNA damage condition by HU depending on FgV2 infection. Colony pictures were taken 5 days after inoculation. (C) Relative colony growth and FgV2 accumulation level of FgV2-infected TF mutants with HU. Asterisk means the FgV2 accumulation level is statistically different (P<0.05) from that in each 0 mM HU condition based on Tukey's test.

In addition to DDR-related TF mutants, abnormal colony morphology was detected in some reactive oxygen species (ROS) sensitive TF mutants like ΔGzWing020 and ΔGzZC303 in FgV1 infection (Yu and Kim, 2021). It was reported that ROS in the host is closely related to virus replication and infection (Hyodo et al., 2017; Lin et al., 2016; Yang et al., 2018). In addition, the expression of some cellular redox regulation genes was significantly changed following mycovirus infections such as Malassezia sympodialis mycovirus 1 and Fusarium graminearum hypovirus 1 (FgHV1), and the p20 of FgHV1 also triggered ROS accumulation in Nicotiana benthamiana (Applen Clancey et al., 2020; Wang et al., 2016). In this regard, we compared phenotypes of TF mutants related to ROS sensitivity to investigate a possible association between TFs and FgV2 infection (Fig. 5A). Among ROS-related TF mutants, some FgV2-infected TF mutants, including ΔGzWing020 and ΔGzZC121, showed fast mycelial growth. These ROS-related TFs in Group 3 showed lower FgV2 dsRNA accumulation levels than WT-VI (Fig. S5).

Fig. 5.

(A) Colony morphology of FgV2-infected TF deletion mutants associated with the oxidative stress response. All colony pictures were taken after 5 days of incubation on CM agar media. Os, sensitive to osmotic stress; Fung, sensitive to fungicide response; CW, sensitive to cell wall stress; pH 4 and pH 11, sensitive responses to pH 4 and pH 11 stress, respectively; DDR, sensitive to DNA damage stress. (B) NBT staining was conducted to detect ROS accumulation in WT and TF mutants related to the oxidative stress response. Colonies were grown in CM agar for 3 days. VF and V2 means virus-free and FgV2-infected, respectively. (C) The relative expression level of superoxide-related genes. The transcript level of FgNoxA, FgNoxB, FgSOD3, Fca7, and Fpx15 was measured by semi-quantitative RT-PCR with 25 cycles (Fpx15) or 30 cycles. EF1α was used as a control (25 cycles). PCR product was visualized on 1% agarose gel with UV. Sizes of amplified target fragments are marked on the right.

ROS are accumulated under abiotic or biotic stress conditions and play essential roles in signaling pathways. Several ROS-responsive genes and transcription factors were identified in fungi, including F. graminearum (Lee et al., 2018). To understand that FgV2 infection causes ROS production and to identify some TFs that might be related to the regulation of ROS accumulation, we assessed nitro blue tetrazolium (NBT) staining for detecting ROS accumulation (Fig. 5B). We observed a slightly higher accumulation of blue formazan precipitates in WT-VF than in WT-VI. Among TF mutants, virus-free ΔGzWing020 and FgV2-infected ΔGzZC121 showed high ROS accumulation levels in NBT staining. To explain the high accumulation of ROS in ΔGzWing020 and ΔGzZC121, we examined the transcript level of some ROS-generating or ROS-scavenging genes by semi-quantitative RT-PCR. FgNoxA and FgNoxB are known to be involved in ROS production, and FgSOD3 is reported to function as a superoxide scavenger in the cytoplasm (Furukawa et al., 2017; Wang et al., 2014). Fca7 and Fpx15 are suggested to be involved in oxidative stress response scavenging hydrogen peroxide (Lee et al., 2018). The expressions of FgNoxA and FgNoxB genes were increased in ΔGzWing020 compared to WT, indicating that excessive ROS production in ΔGzWing020 might be related to the induction of FgNoxA and FgNoxB (Fig. 5C). In Fig. 5B, we observed that ROS accumulation was significantly increased in ΔGzWing020. However, it was greatly decreased upon FgV2 infection. When we examined gene expression of ROS scavenging genes, FgSOD3, Fca7, and Fpx15 were noticeably induced in FgV2-infected ΔGzWing020. This result indicated that deletion of GzWing020 might affect the regulation of ROS production, and the increased expression of ROS scavenging-related genes led to reduced ROS accumulation in FgV2-infected ΔGzWing020. In the case of ΔGzZC121, ROS production-related genes and ROS scavenging genes showed increased gene expression following FgV2 infection, although a high accumulation of ROS was detected in FgV2-infected ΔGzZC121. It might be explained as the possible involvement of GzZC121 in the ROS-scavenging pathway.

3.6. Protein-protein interactions of TFs related to FgV2 infection

To determine the association among TFs related to FgV2 infection, we constructed protein-protein interaction (PPI) network of TFs combining transcriptome data that analyzed changes in gene expression (log2 fold changes>0.5) followed by FgV2 infection (Lee et al., 2014). Among 384 TFs, 88 TFs constructed the largest protein-protein direct interaction network around GzbZIP015 (Fig. 6). Among 88 TFs, 6 TFs were in Group 1, 15 TFs were in Group 3, and 3 TFs were FgV2 non-transmissible. Functional annotation was conducted using FungiFun 2.2.8 with FunCat classification ontology. As a result, most TFs were expectably involved in transcription and nucleic acid binding, and other functional categories like ‘Metabolism’ were also detected. Among 88 TFs, 16 TFs were involved in metabolism, and most were directly connected around GzbZIP015. Gene expression of these metabolism-related TFs was generally up-regulated by FgV2 infection. A total of 16 TFs were annotated with ‘Biogenesis of cellular component’ or ‘Cell cycle and DNA processing’ and mostly showed reduced gene expression following FgV2 infection. These TFs also constructed a small directly connected network. Additionally, ‘Cell rescue, defense, and virulence’ was annotated to several TFs. These TFs were scattered across the PPI network but closely existed with other TFs in ‘Metabolism’, ‘Biogenesis of cellular component’ or ‘Cell cycle and DNA processing. To compare with FgV1-host interaction, we also generated a PPI network of TFs combined with transcriptome in response to FgV1 infection (Fig. S6). In FgV1, the predicted interactome consists of 219 interactions with 98 TFs. Among them, 48 TFs were demonstrated in both fungal host PPI networks in response to FgV1- and FgV2 infection. Similar to FgV2 infection, the main hubs of predicted PPI networks included GzbZIP015 annotated by ‘metabolism’. More than 3 times TF interactions were shown in ‘Biogenesis of cellular component’, ‘Cell cycle and DNA processing’, and ‘Cell rescue, defense, and virulence’ compared with the FgV2 network. Moreover, we could predict additional interactions among TFs annotated to ‘Protein synthesis’ and ‘Cell fate’ which are not clearly shown in TF-TF networks in response to FgV2 infection.

Fig. 6.

Protein-protein interaction of TFs differentially expressed upon FgV2 infection. A total of 384 TFs (log2 fold changes>0.5) were subjected to the prediction of PPI using the Cytoscape STRING program, and functional description was conducted using FungiFun 2.2.8 with FunCat classification ontology. Except for transcription, the main functional categories were indicated using each color. Directly connected TFs in the same functional category were linked with a colored edge. TFs with no direct connection with other TFs in the same functional category were indicated with a colored border.

4. Discussion

Hosts turn on the defense mechanisms against virus infection by regulating many transcription factors (Denard et al., 2011; Huh et al., 2012). Viruses also interfere with and control host transcription factors against the host defense system and replicate using pro-viral host factors (Tan et al., 2022; Zhang et al., 2020a). In this regard, identifying TFs related to virus infection is significant to understanding virus-host interactions. Here, we transferred FgV2 to the F. graminearum TF gene deletion mutant library to identify TFs involved in the FgV2–host interaction. As shown in Fig. 2, we observed the inverse correlation between mycelial growth and FgV2 dsRNA accumulation level in FgV2-infected TF mutants that showed similar observation in FgV1-infected TF mutants previously (Yu and Kim, 2021). These results suggested the possibility that some of the TFs in Group 1 are anti-viral factors and some of the TFs in Group 3 are pro-viral factors. However, these putative pro- or anti-viral factors are quite different between FgV1- or FgV2-infected strains (Fig. S2, (Yu and Kim, 2021)). Previously we confirmed that virus infections on F. graminearum pH-1 affect fungal phenotypes differently, including sexual development, conidiation, conidial morphology, and trichothecene production (Lee et al., 2014). We assumed that these differences in phenotypic changes were caused by the involvement of different host factors, pathways, and processes in FgV1 or FgV2 infection, although both FgV1 and FgV2 mediate hypovirulence on their host. In addition, several studies observed that the same host transcription factor acts oppositely upon different virus infections. For example, the signal transducer and activator of transcription 3 (STAT3) inhibit herpes simplex virus-1 by regulating the immune system (Hsia et al., 2017). On the other hand, STAT3 acts as a proviral host factor in hepatitis C virus infection by promoting microtubule polymerization (McCartney et al., 2013). Comparing the common or unique effects of FgV1- or FgV2-infected TF deletion mutants in further investigations will provide insight into interactions between FgV1 and FgV2 in F. graminearum.

We observed DI-RNAs in FgV2-infected WT and several TF mutants. Diverse effects of defective viral genomes have been studied, including interference and viral evolution (Vignuzzi and López, 2019). We evaluated the effect of DI-RNAs of FgV2 on its parental virus and fungal colony (Fig. 3). DI-RNAs generally interfered with FgV2 gRNA accumulation, resulting in an increased mycelial growth rate. Similarly, DI-RNA of Rosellinia necatrix partitivirus 6 (RnPV6) inhibited genomic dsRNA accumulation and attenuated the colony growth reduction by RnPV6 infection in Cryphonectria parasitica (Chiba et al., 2016). In addition, we observed that the presence of DI-RNAs harmed FgV2 transmission through conidiation (Fig. 3E). Unlike FgV2, vertical transmission of Trichoderma harzianum hypovirus 1 (ThHV1) and its d-RNA (ThHV1-S) showed a higher vertical transmission rate when both ThHV1 and ThHV1-S existed together in Trichoderma harzianum (You et al., 2019).

In C. parasitica, it was reported that deletion in DI-RNAs of Cryphonectria hypovirus 1 (CHV1) is matched with the region where virus-derived small RNA (vsRNA) was infrequent and DI-RNAs were not generated in Δdcl2 or Δago2 mutant, suggesting a possible association of RNAi pathway with DI-RNA generation (Nuss, 2011). To confirm this possible association, we compared the transcript level of FgDICER-2 and FgAGO-1 in FgV2-infected WT and ΔGzZC138 depending on the DI-RNAs presence (Fig. S7). However, the gene expression level of FgDICER-2 and FgAGO-1 was relatively low in DI-RNA-harboring colonies compared to DI-RNA-free, presumably because of a lower FgV2 accumulation level. Although the specific mechanisms for generating DI-RNA are not clear, we expected that the deletion of some TF genes led to higher accumulation/generation of DI-RNA. It was found that most of the DI-RNAs in these mutants and WT-VI were generated from FgV2 RNA3 without a C-terminal region (Figs. 3D and S4). Similarly, it was reported that detected DI-RNAs of RnPV6, which has two segmented RNAs, were only generated from RNA1 encoding RdRp (Chiba et al., 2016). Considering the higher nucleotide similarity between FgV2 and FgV-ch9, we expected that ORF3 of FgV2 encodes CP and is related to the symptom severity of virus infection (Bormann et al., 2018; Darissa et al., 2011). It is uncertain that DI-RNAs derived from RNA3 function as a symptom development factor and eventually inhibit virus replication, so further investigations are required.

It was revealed that the Spt–Ada–Gcn5 acetyltransferase (SAGA) complex, one of the transcriptional co-activator, mediates transcriptional induction of dcl2 and agl2 in C. parasitica by CHV1 infection (Andika et al., 2017). However, the specific mechanisms of transcriptional regulation of RNAi-related genes by mycovirus infection in fungi are not well studied. In the previous study, we confirmed that FgV1 pORF2 has a role as a suppressor of RNAi response by binding to the promoter region of RNAi-related genes and inhibiting the induction of those genes (Yu et al., 2020). We assumed that some transcription factors might be required to induce FgDICERs and FgAGOs gene expression in response to virus infection. In this study, using the FgV2-infected TF deletion mutant library, we found that FgDICER-2 and FgAGO-1 were not fully induced in FgV2-infected ΔGzZC137 despite the high accumulation of FgV2, suggesting a possible role of GzZC137 in the induction of these genes.

Conversely, FgV2-infected ΔGzGATA007 and ΔGzMIZ001 showed significant induction of FgDICER-2 and FgAGO-1 with the low accumulation of FgV2 gRNA. In the case of ΔGzGATA007, relatively high accumulated DI-RNAs can be one reason for the increased induction of FgDICER-2 and FgAGO-1 despite the small amount of gRNA because DI-RNAs can trigger antiviral immunity (Fig. S3; Vignuzzi and López, 2019). Since GzZC137, GzGATA007, and GzMIZ001 are functionally unknown genes in fungi, including F. graminearum, further studies on identifying their cellular functions and roles in the transcriptional regulation of RNAi components are necessary.

Many studies have shown that some RNA viruses induce and harness host DDR pathways and components, although they generally replicate in cytoplasm exclusively (Ryan et al., 2016). We observed that the mycelial growth of ΔGzGATA007 was significantly inhibited upon FgV2 infection, and FgV2 accumulation was significantly increased in ΔGzGATA007 under DNA damage stress condition by HU. GzGATA007 is the homolog of SreA, which is involved in iron uptake regulation in A. nidulans (Schrettl et al., 2008). Specific resistance mechanisms against HU by GzGATA007 deletion have not been discovered yet. Still, it is suggested that FgV2 might impose additional DNA damage stress causing alteration of the HU resistance mechanism or taking benefits from DNA damage resistance mechanisms. It was reported that a considerable reduction in nuclear localization and an increase in the cytoplasm and cytoplasmic foci by protein re-localization occur in response to genotoxic chemicals in yeast (Tkach et al., 2012). Therefore, it would also be required to consider protein re-localization in DDR pathway-related components following FgV2 infection to identify the association of FgV2 infection with DNA damage in F. graminearum. Recently, it was reported that many genes associated with DNA replication and DDR process in Sclerotinia sclerotiorum were up-regulated by Sclerotinia sclerotiorum hypovirulence-associated DNA virus 1 infection (Qu et al., 2021). This suggests that mycovirus is also closely connected to the host DDR pathway, and this area will be valuable to investigate for understanding interactions between mycovirus and fungi.

A relationship between ROS and ROS regulating pathways with virus replication has been studied in other host systems (Chen et al., 2012; Hyodo et al., 2017; Yang et al., 2018). In phytopathogenic fungi, the role of ROS was studied mainly focused on fungal development and interactions with the plant (Heller and Tudzynski, 2011; Zhang et al., 2020b). Although many mycoviruses have been identified, little is known about the relationship between hypovirulence-associated mycovirus and ROS in the host cell. In this study, we found that FgV2 accumulation was decreased in TF mutants in which ROS was highly accumulated (Figs. 5B, S5). These mutants showed differential expression patterns of FgNoxA, FgNoxB, FgSOD3, Fca7, and Fpx15 compared to WT. It seems that GzWing020 inhibits FgNoxA and FgNoxB expressions. Investigated ROS-related genes were also differentially expressed in ΔGzZC121, suggesting the possible role of GzZC121 in ROS regulation. Although some viruses induce ROS generation for their replication (Hyodo et al., 2017; Lin et al., 2016), increased ROS level seems detrimental to FgV2 replication from the results of this study. It was reported that ROS accumulation has adverse effects on virus replication by disrupting viral replication complexes or activating host defense mechanisms at the early infection stage (Choi et al., 2004; Liu et al., 2021). The expression level of FgSOD3, Fca7, and Fpx15, ROS scavenging enzymes, was noticeably increased in ΔGzWing020 and ΔGzZC121 following FgV2 infection. Although the specific induction mechanism of these genes is unknown, this result might be one of the adaptive processes of FgV2 for survival in a ROS-excessive environment.

We found several TF deletion mutants that FgV2 could transmit via protoplast fusion, not hyphal fusion. We assumed that one of the reasons might be the deletion of TF affected during hyphal fusion processes between virus-free ΔTF and WT-VI. It was reported that Sclerotinia sclerotiorum mycoreovirus 4 (SsMYRV4) suppresses ROS generation and regulates vegetative incompatibility in Sclerotinia sclerotiorum, resulting in heterologous mycovirus transmission by hyphal fusion (Wu et al., 2017). Among the TF mutants in which FgV2 was not transmitted by hyphal fusion, such as ΔGzWing020 and ΔGzZC121, high accumulation of ROS caused poor efficiency of FgV2 transmission by hyphal fusion. On the other hand, it is possible that they contained undetectable FgV2 viral RNA levels after hyphal fusion normally occurred. We confirmed that FgV2-infected ΔGzNH001 and ΔGzOB007 showed very similar morphology with virus-free colonies when we obtained virus-infected mutants by protoplast fusion (Fig. S1). Compared to FgV1 non-transmissible TF mutants, we found 9 TF deletion mutants, including ΔGzZC316, ΔGzZC301, and ΔGzZC060, to which both FgV1 and FgV2 were not transmitted (Table S1). Among these TFs, GzZC232 encodes the orthologue of C. parasitica PRO1, which is required for the stable maintenance of CHV1-EP713 in C. parasitica (Sun et al., 2009a). Because FgV1 and FgV2 infection was not detected at all in these TF mutants, it is plausible that some of these TFs are essential for the survival of FgV1 and FgV2 in the host cell or negative regulator in the anti-viral pathway, which deletion can induce a strong anti-viral response in initial virus infection stage. We confirmed the gene expression of virus non-transmissible TF genes in WT following FgV2 infection (Table S4). Unlike FgV1, most TF genes were not differentially expressed in FgV2 infection except GzZC030 and GzZC316 (Yu and Kim, 2021). Following FgV1 or FgV2 infection, GzZC030 and GzZC316 expression showed significant changes. Further studies will be required to understand the functions of these TFs in unknown regulatory networks and the effect of horizontal virus transfer.

To understand the changes in TF networks in response to FgV2 infection, we predicted the PPI of TFs in F. graminearum combined with transcriptome data (Fig. 7). In this PPI network, we observed many TFs constituted a direct connecting network around GzbZIP015, which is the ortholog of GCN4 in the Saccharomyces cerevisiae or CPC1 in C. parasitica (Hinnebusch, 2005; Timpner et al., 2013). Among downstream TFs from the GzbZIP015, deletion mutants of several TFs, including GzRFX1, GzLam002, and GzGATA007, were grouped into Group 3. It suggested that FgV2 might interact with GzbZIP015-associated host factors or signaling pathways for virus replication or pathogenesis. GzLam002 is a putative ortholog of Saccharomyces cerevisiae multiprotein-bridging factor 1 (Mbf1) and functions as a co-activator of GCN4 by bridging between GCN4 and TATA-binding protein (TBP) in response to nutritional stress (Takemaru et al., 1998). Deletion of GzLam002 affects virus accumulation or virus-associated phenotype in both FgV1 and FgV2. We assumed that interaction between GzbZIP015 and GzLam002 also existed in F. graminearum, and their functions might be required for virus accumulation. Compared to the FgV2-responsive PPI network in the host, the FgV1-responsive PPI network showed notable features in GzMADS003- and GzGATA007-related PPI interactions rather than GzbZIP015-associated PPI interactions (Fig. S6). These results also indicate that different host factors are involved in FgV1 and FgV2 infections, and these differences cause distinguishable phenotypes in the fungal host.

Taken together, our studies help understand interactions between FgV2 and F. graminearum and provide insights to find noble regulatory genes/pathways involved in stress response under biotic stress conditions.

Author contributions

J.Y. and K-H.K. designed the experiments; G.K. and J.Y. performed the experiments and analyzed the data; G.K. drafted the first manuscript; J.Y. and K-H.K. supervised the experiments and edited the manuscript.

Declaration of Competing Interest

The authors declare that they have no conflict of interest. This article does not contain any studies performed by any authors with human participants or animals.

Data availability

• No data was used for the research described in the article.