Genome-wide exonic small interference RNA-mediated gene silencing regulates sexual reproduction in the homothallic fungus Fusarium graminearum

Abstract

Various ascomycete fungi possess sex-specific molecular mechanisms, such as repeat-induced point mutations, meiotic silencing by unpaired DNA, and unusual adenosine-to-inosine RNA editing, for genome defense or gene regulation. Using a combined analysis of functional genetics and deep sequencing of small noncoding RNA (sRNA), mRNA, and the degradome, we found that the sex-specifically induced exonic small interference RNA (ex-siRNA)-mediated RNA interference (RNAi) mechanism has an important role in fine-tuning the transcriptome during ascospore formation in the head blight fungus Fusarium graminearum. Approximately one-third of the total sRNAs were produced from the gene region, and sRNAs with an antisense direction or 5'-U were involved in post-transcriptional gene regulation by reducing the stability of the corresponding gene transcripts. Although both Dicers and Argonautes partially share their functions, the sex-specific RNAi pathway is primarily mediated by FgDicer1 and FgAgo2, while the constitutively expressed RNAi components FgDicer2 and FgAgo1 are responsible for hairpin-induced RNAi. Based on our results, we concluded that F. graminearum primarily utilizes ex-siRNA-mediated RNAi for ascosporogenesis but not for genome defenses and other developmental stages. Each fungal species appears to have evolved RNAi-based gene regulation for specific developmental stages or stress responses. This study provides new insights into the regulatory role of sRNAs in fungi and other lower eukaryotes.

Fig 1. Characterization of Dicers and Argonautes in F. graminearum.

(A) Domain architectures of F. graminearum Dicer and Argonaute proteins along with those of representative orthologs. Domains were predicted using SMART [38]. Conservation analysis of the Asp-Asp-His (DDH) motif that comprises the active sites of Argonautes. The DDH motif is conserved in FgAgo2, but the third histidine residue is replaced by an aspartate in FgAgo1. The amino acid sequences of orthologs: Homo sapiens Dicer1 (NP_803187), Arabidopsis thaliana Dcl2 (NP_566199), Schizosaccharomyces pombe Dcl1 (Q09884), F. graminearum FgDicer1 (Gene ID: FGSG_09025), FgDicer2 (Gene ID: FGSG_04408), H. sapiens Ago2 (NP_036286), A. thaliana Ago1 (NP_849784), S. pombe Ago1 (NP_587782), F. graminearum FgAgo1 (Gene ID: FGSG_16976), and F. graminearum FgAgo2 (Gene ID: FGSG_00348). (B) Expression profiles of FgDICER1, FgDICER2, FgAGO1, and FgAGO2 in the F. graminearum wild-type strain during vegetative and sexual development. Transcript levels were analyzed via qRT-PCR during the vegetative (V3 and V5, 3 and 5 days after inoculation, respectively) and sexual stages (S3, S5 and S7, 3, 5, and 7 days after sexual induction, respectively) on carrot agar. The transcript level of the gene at the 3-day vegetative stage (V3) was arbitrarily set to 1, and this value was used for comparison to other periods. (C) Forcible ascospore discharge. A semi-circular agar block covered with perithecia was placed on a coverslip. Images were collected 48 h after the assay was initiated. White cloudy material (indicated with an arrow) represents discharged ascospores. (D) Number of discharged ascospores. Discharged ascospores were corrected for one week from the 7-day-old sexually induced cultures. (E) Asci rosettes. Imaging was performed 8 days after sexual induction. Red arrows indicate asci with defective ascospore delimitation. Scale bar = 20 μm.

Fig 2. Transcriptome analysis of RNAi component mutants.

Venn diagrams illustrating the overlap between upregulated (A) and downregulated (B) genes in the RNAi component mutants compared to those of the wild type. DEGs were identified as genes showing a greater than 3-fold change in transcript levels compared to those of the wild type. (C) Correlation analyses of transcriptomes of Fgdicer1 Fgdicer2 and Fgago1 Fgago2. (D) Venn diagrams illustrating the overlap between DEGs of the Fgdicer1 Fgdicer2 and Fgago1 Fgago2 mutants. (E) Correlation analyses of transcriptomes of Fgdicer1 Fgdicer2 and Fgdicer1, Fgdicer1 Fgdicer2 and Fgdicer2, Fgago1 Fgago2 and Fgago1, and Fgago1 Fgago2 and Fgago2.

Fig 3. Characterization of DEGs in RNAi-deficient mutants.

(A) Expression profiles of clustered groups including the mating-type genes. Fuzzy clustering categorized total genes into 10 groups depending on their expression profiles during sexual development (0–5 day after sexual induction). The genes included in groups 2 and 9 showed similar expression patterns as those of the mating-type genes (MAT1-1-1, MAT1-1-2, MAT1-1-3, and MAT1-2-1) during sexual reproduction. Transcriptome data during sexual development were obtained from a previous study [8] and re-analyzed for this study. (B) Gene Ontology (GO) enrichment network of the upregulated genes in the RNAi-deficient mutants. (C) GO enrichment network of the downregulated genes in the RNAi-deficient mutants. GO terms were statistically analyzed using GOstats [40] and visualized using REVIGO [41].

Fig 4. Characterization of sRNAs.

(A) Nucleotide preference of 5′ end and size distribution of sRNAs produced by F. graminearum strains. (B) Analysis of 5′ end nucleotide preference of sRNAs produced by F. graminearum strains. (C) Relative abundance of sRNAs mapped to various genomic features produced by F. graminearum strains. (D) Absolute abundance of sRNAs mapped to various genomic features produced by F. graminearum strains. -, sRNAs with antisense direction; +, sRNAs with sense direction; 5′-U, sRNAs with 5′ uracil (22–25 nt).

Fig 5. Characterization of sRNA-producing genes.

(A) Number of genes that produce sRNAs depending on F. graminearum strains. (B) Number of genes that produce sRNAs with 5′-U depending on F. graminearum strains. (C) Correlation analyses of transcriptomes of Fgdicer1 Fgdicer2 and Fgago1 Fgago2 compared to that of the wild type depending on sRNA counts. The log₂ ratio of transcript abundance in Fgdicer1 Fgdicer2 versus wild-type (x axis) and Fgago1 Fgago2 versus wild-type (y axis) is plotted. Colors indicate the sRNA density (reads per kilobase). (D) Correlation analyses of sRNA counts and transcript abundance. Gene numbers with corresponding log₂ ratio of transcript abundance in Fgdicer1 Fgdicer2 or Fgago1 Fgago2 versus the wild-type strain Z-3639 were counted. Most genes producing antisense sRNAs more than 1000 counts per kilobase (red graphs) were positively regulated in Fgdicer1 Fgdicer2 and Fgago1 Fgago2 compared to those in the wild type. Colors indicate the sRNA density denoted in Fig 5C.

Fig 6. Detection of ex-siRNAs.

Quantification of ex-siRNA candidates was performed using stem-loop RT-PCR assays. Each 100 ng of small RNA-enriched RNA samples was used for reverse transcription reactions, and the images were obtained from 25–30 cycles of PCR using 10% acrylamide gel.

Fig 7. Representative IGV images of genes possibly regulated by ex-siRNA.

(A) Aligned sRNA-seq and mRNA-seq results of F. graminearum strains were visualized using IGV. Red stars indicate selected ex-siRNA for RT-PCR. (B) Relative transcript abundances of genes in the F. graminearum strains during sexual development. Transcript levels were analyzed via qRT-PCR 5 days after sexual induction. The transcript level of the wild type was arbitrarily set to 1.

Fig 8. Degradome analyses.

(A) Histogram displaying the 5′ positions of degradome tags from two wild-type libraries relative to normalized transcript positions. The protein-coding sequence of the transcript is exhibited from 0 to 100. (B) Frequency of degradome tags mapped to various genomic features produced by the wild-type strain. (C) Correlation analyses between wild-type degradome counts and transcript abundances of wild-type or Fgdicer1 Fgdicer2 strains. (D) Correlation analyses of transcriptomes of Fgdicer1 Fgdicer2 and Fgago1 Fgago2 compared to that of the wild type. Colors indicate the degradome tag density (reads per kilobase). (E) Average degradome counts depending on percentile ranks of antisense sRNA counts. Percentile ranks of genes were assigned based on their antisense sRNA counts. (F) Average transcript abundance of Fgdicer1 Fgdicer2 compared to the wild type depending on the percentile ranks of the degradome tags/transcript counts ratio. (G) Number of sRNA-producing genes corresponding to the percentile ranks of the degradome tags/transcript counts ratio. Numbers of sRNA-producing genes were divided into four groups depending on the transcript abundance of Fgdicer1 Fgdicer2 compared to the wild type.