Differential Contribution of RNA Interference Components in Response to Distinct Fusarium graminearum Virus Infections

Abstract

The mechanisms of RNA interference (RNAi) as a defense response against viruses remain unclear in many plant-pathogenic fungi. In this study, we used reverse genetics and virus-derived small RNA profiling to investigate the contributions of RNAi components to the antiviral response against Fusarium graminearum viruses 1 to 3 (FgV1, -2, and -3). Real-time reverse transcription-quantitative PCR (qRT-PCR) indicated that infection of Fusarium graminearum by FgV1, -2, or -3 differentially induces the gene expression of RNAi components in F. graminearum Transcripts of the DICER-2 and AGO-1 genes of F. graminearum (FgDICER-2 and FgAGO-1) accumulated at lower levels following FgV1 infection than following FgV2 or FgV3 infection. We constructed gene disruption and overexpression mutants for each of the Argonaute and dicer genes and for two RNA-dependent RNA polymerase (RdRP) genes and generated virus-infected strains of each mutant. Interestingly, mycelial growth was significantly faster for the FgV1-infected FgAGO-1 overexpression mutant than for the FgV1-infected wild type, while neither FgV2 nor FgV3 infection altered the colony morphology of the gene deletion and overexpression mutants. FgV1 RNA accumulation was significantly decreased in the FgAGO-1 overexpression mutant. Furthermore, the levels of induction of FgAGO-1, FgDICER-2, and some of the FgRdRP genes caused by FgV2 and FgV3 infection were similar to those caused by hairpin RNA-induced gene silencing. Using small RNA sequencing analysis, we documented different patterns of virus-derived small interfering RNA (vsiRNA) production in strains infected with FgV1, -2, and -3. Our results suggest that the Argonaute protein encoded by FgAGO-1 is required for RNAi in F. graminearum, that FgAGO-1 induction differs in response to FgV1, -2, and -3, and that FgAGO-1 might contribute to the accumulation of vsiRNAs in FgV1-infected F. graminearumIMPORTANCE To increase our understanding of how RNAi components in Fusarium graminearum react to mycovirus infections, we characterized the role(s) of RNAi components involved in the antiviral defense response against Fusarium graminearum viruses (FgVs). We observed differences in the levels of induction of RNA silencing-related genes, including FgDICER-2 and FgAGO-1, in response to infection by three different FgVs. FgAGO-1 can efficiently induce a robust RNAi response against FgV1 infection, but FgDICER genes might be relatively redundant to FgAGO-1 with respect to antiviral defense. However, the contribution of this gene in the response to the other FgV infections might be small. Compared to previous studies of Cryphonectria parasitica, which showed dicer-like protein 2 and Argonaute-like protein 2 to be important in antiviral RNA silencing, our results showed that F. graminearum developed a more complex and robust RNA silencing system against mycoviruses and that FgDICER-1 and FgDICER-2 and FgAGO-1 and FgAGO-2 had redundant roles in antiviral RNA silencing.

Keywords: Argonaute; Fusarium graminearum virus; RNA silencing; antiviral response; mycovirus.

FIG 1

Colony morphologies and FgV1 RNA accumulation levels of the Fusarium graminearum RNA silencing mutant strains. (A) Generation of mutants in F. graminearum. Gene deletion (top), complementation (middle), and overexpression (bottom) mutants were generated in this study. Schematic representations of the homologous gene recombination strategies used to generate the RNA silencing-related F. graminearum mutants (left) and results of Southern blot hybridization of the mutants (right) are shown. The ORFs of the target genes were fused with the hygromycin resistance cassette to generate complementary strains. The promoter was replaced with the elongation factor 1α promoter (P_EF1α) in the overexpression strain. All DNA probes and restriction enzymes used for constructing each mutant strain are shown on the right, and the expected DNA sizes indicated. A ³²P-labeled DNA fragment of the 5′-flanking or 3′-flanking region of the target gene was used as a probe for Southern blot hybridization. The sizes of the DNA bands are indicated to the left of the Southern blot images. Lanes 1, wild-type (WT) strain PH-1; lanes 2 to 4, different biological replicates of the indicated single gene deletion mutant used in this study; lane(s) n, nonhomologous (ectopic) insertion that was not selected. (B) Colony morphologies of virus-free (left) and Fusarium graminearum virus 1 (FgV1)-infected (right) RNA silencing gene deletion, complementation, and overexpression mutant strains. All cultures were photographed after 5 days on CM. (C) Semiquantification of viral dsRNA accumulation levels of the wild type and different mutant strains. Shown is the viral dsRNA accumulation level of each RNA silencing mutant relative to that of PH-1/FgV1, which was set at one as determined with ImageJ software. Values were calculated using the results from three independent biological replicates, and error bars indicate standard deviations (SD). (D and E) Quantification of plus-strand (D) and minus-strand (E) viral RNA accumulation at 120 h postinoculation (hpi) using real-time reverse transcription-quantitative RT-PCR (qRT-PCR). EF1α and UBH gene transcripts were used as internal controls. Mean values (± SD) from two biological replicates and at least three replicated experiments are shown. (C to E) Mean values with different numbers of asterisks are significantly different (P < 0.05) from each other based on Tukey's test. KO, gene deletion (knockout) mutant; COM, complementation mutant; OE, overexpression mutant.

FIG 2

Analysis of the double gene deletion mutants. (A) Colony morphologies of ΔFgDICER-1/FgDICER-2 and ΔFgAGO-1/FgAGO-2 double gene deletion virus-free and virus-infected mutants. Cultures were photographed after 5 days on CM. (B) Accumulation of FgV1 RNA in double gene deletion mutants at 120 hpi according to qRT-PCR. Means with an asterisk are significantly different (P < 0.05) from the mean of the virus-infected wild type (PH-1/FgV1) based on Tukey's test. (C) Agarose gel (1%) analysis of the dsRNA accumulation for PH-1/FgV1 and double gene deletion mutants. A 10-μg quantity of total RNA per sample was treated with DNaseI and S1 nuclease. The largest band in the FgV1-infected samples represents the full-length viral dsRNA (6.6 kb); smaller bands indicate internally deleted forms of viral dsRNA. Lane M, lambda DNA digested with HindIII.

FIG 3

Colony morphologies and FgV1 RNA accumulation levels in cross-double gene deletion mutants of F. graminearum. (A) Colony morphologies of ΔFgDICER-1/FgAGO-1, ΔFgDICER-2/FgAGO-1, ΔFgDICER-1/FgAGO-2, and ΔFgDICER-2/FgAGO-2 mutants. Cultures were photographed after 5 days on CM. (B) Accumulation of FgV1 RNA in mutant strains at 120 hpi according to qRT-PCR. Mean values with different numbers of asterisks are significantly different (P < 0.05) from each other based on Tukey's test.

FIG 4

Accumulation of FgDICER-1, FgDICER-2, and FgAGO-1 gene transcripts in response to FgV1 infection. The transcript levels of FgDICER-1, FgDICER-2, and FgAGO-1 in the FgV1-infected wild type (PH-1/FgV1) and the FgV1-infected mutants were determined by qRT-PCR. Mean values (± SD) are shown. Mean values with different numbers of asterisks are significantly different (P < 0.05) from each other based on Tukey's test. OE, overexpression.

FIG 5

Functional analysis of dicer and Argonaute in hairpin RNA-induced gene silencing. (A) Diagram of transformation of plasmids used to study GFP silencing. The pSKGen vector was under the control of the EF1 alpha promoter, and the pGFP-SA construct was under the control of the isocitrate lyase (ICL) promoter. A HincII fragment of the β-glucuronidase (GUS) gene was used as a spacer. Segments used as probes in Southern blot hybridization analysis are indicated by bars. GFP, transformant that only had the pSKGen construct; SA, transformant with the pSA construct. (B) Colony morphologies of virus-free GFP silencing-related transformants and FgV1-infected mutant strains. Cultures were photographed after 5 days on CM. GFP+SA, GFP and pSA were transformed together into the wild type. (C) GFP expression in mutant strains. Mycelia of GFP transformants and GFP-silenced transformants (GFP+SA) were observed with a microscope at 2 days postincubation in CM. DIC, differential interference contrast. Scale bar = 20 μm. (D) Accumulation of FgDICER-2 and FgAGO-1 gene transcripts in GFP expression or hairpin dsRNA-producing mutants. Accumulation levels of FgDICER-2 and FgAGO-1 in all virus-free and FgV1-infected mutants relative to the levels in the virus-free wild type (PH-1), which were set to one, are shown. EF1α and UBH gene transcripts were used as the internal controls. (E and F) Accumulation of FgV1 (E) and FgV2 (F) RNA in mutant strains. qRT-PCR was used to quantify FgV1 and FgV2 RNA1 at 120 hpi. (G) Confirmation of accumulation of FgDICER-1 and FgRdRP1 to -5 in hairpin RNA-induced gene silencing mutants. Gene expression levels of FgDICER-1 and five FgRdRP genes were analyzed by qRT-PCR at 120 hpi. (D to G) Mean values (± SD) from at least two biological replicates and three independent experiments are shown. Mean values with different numbers of asterisks are significantly different (P < 0.05) from each other based on Tukey's test.

FIG 6

Accumulation of FgV RNA in virus-infected RNA silencing gene mutant strains of F. graminearum. (A) Colony morphologies of mutant strains infected with Fusarium graminearum virus 2 (FgV2). (B) Quantification of FgV2 RNA1 at 120 hpi using qRT-PCR. (C) Colony morphologies of mutant strains infected with FgV3. (D) Quantification of FgV3 RNA at 120 hpi using qRT-PCR. (A and C) All cultures were photographed after 5 days on CM. (B and D) Mean values (± SD) of two biological replicates and at least three replicated experiments are shown. EF1α and UBH gene transcripts were used as internal controls. Mean values with different numbers of asterisks are significantly different (P < 0.05) from each other based on Tukey's test. (E) Colony morphologies of ΔFgDICER-1/FgDICER-2 or ΔFgAGO-1/FgAGO-2 double gene deletion virus-free and virus-infected mutants. Cultures were photographed after 5 days on CM. (F) Agarose gel (1%) analysis of dsRNA accumulation in PH-1/FgV2, PH-1/FgV3, and double gene deletion mutants. A 20-μg quantity of total RNA per sample was treated with DNaseI and S1 nuclease. Lane M1, 1-kb ladder; lane M2, lambda DNA digested with HindIII.

FIG 7

Identification of vsiRNAs by small RNA sequencing. (A) Numbers of vsiRNAs identified (y axis) based on size (number of nucleotides; x axis). (B) Percentages of vsiRNAs located on sense and antisense strands of each virus. (C) Percentages of identified vsiRNAs located on individual RNA segments of FgV2.

FIG 8

Profiling of viral small interfering RNAs (vsiRNAs) on FgV1, -2, and -3 genomes. The identified vsiRNAs induced by FgV1, FgV2, or FgV3 are mapped on the corresponding virus genome. (A to C) Distributions of vsiRNAs mapped on the FgV1, FgV2, and FgV3 genomes. FgV3 has only one RNA segment, while FgV2 has five RNA segments. The genome organization for each virus genome is depicted. Blue and red bars indicate vsiRNAs located on the sense strand (+) and the antisense strand (−) of the corresponding virus genome, as observed by stem-loop RT-PCR.

FIG 9

Detection of FgV1 vsiRNAs. Stem-loop RT-PCR of vsiRNAs of FgV1 in the 5′-end untranslated region (UTR) (A) and internal region (B). A 10-ng quantity per cDNA sample was used for PCR amplification. The numbers of PCR cycles are indicated to the right. Three independent biological replicates were used for this experiment, and two samples from each were visualized on a 4% agarose gel.