Fungal Argonaute proteins act in bidirectional cross-kingdom RNA interference during plant infection

Abstract

Argonaute (AGO) proteins bind to small RNAs to induce RNA interference (RNAi), a conserved gene regulatory mechanism in animal, plant, and fungal kingdoms. Small RNAs of the fungal plant pathogen Botrytis cinerea were previously shown to translocate into plant cells and to bind to the host AGO, which induced cross-kingdom RNAi to promote infection. However, the role of pathogen AGOs during host infection stayed elusive. In this study, we revealed that members of fungal plant pathogen B. cinerea BcAGO family contribute to plant infection. BcAGO1 binds to both fungal and plant small RNAs during infection and acts in bidirectional cross-kingdom RNAi, from fungus to plant and vice versa. BcAGO2 also binds fungal and plant small RNAs but acts independent from BcAGO1 by regulating distinct genes. Nevertheless, BcAGO2 is important for infection, as it is required for effective pathogen small RNA delivery into host cells and fungal induced cross-kingdom RNAi. Providing these mechanistic insights of pathogen AGOs promises to improve RNAi-based crop protection strategies.

Keywords: Argonaute; cross-kingdom RNA interference; plant–fungal interaction; small RNA.

Fig. 1.

Small RNA profiling of Botrytis cinerea Δbcago ko mutants. (A) Stem-loop RT-PCR of BcsRNAs known to induce cross-kingdom RNAi in plants. Samples were from B. cinerea strains grown in axenic culture or infected tomato (Sl-infected) condition. BcTubulin mRNA was used as an internal control. M: 1 kb DNA ladder. (B) Size profiles of total BcsRNA reads detected in B. cinerea WT and Δbcago ko mutants in two biological replicates by Illumina deep sequencing. (C) Schematic representation of the definition for BcsRNA loci, using a B. cinerea WT small RNA sequencing dataset. (D) Absolute numbers of 21 to 22 nt size-specific BcsRNA loci overlapping with different annotated regions in the B. cinerea genome in the context of retrotransposons (RT). (E) Heat map and cluster analysis of differentially expressed BcsRNA loci comparing B. cinerea WT and Δbcago ko mutants.

Fig. 2.

BcAGO1 and BcAGO2 are required for fungal-induced cross-kingdom RNAi. (A) Fluorescence microscopy images from GFP reporter plant seedlings at different time points. A 5 μl drop of a 5 × 10⁵/ml conidiospore suspension was placed at the center of the leaf before placing a glass covering slip on the top that dispersed the conidiospore suspension over the entire leaf surface. (Scale bar, 1 mm.) (B) GFP quantification in infected seedling leaves from 0 to 42 hpi. For normalization, GFP fluorescence signal intensity at different time points (I_t) was subtracted with the initial GFP signal intensity (I₀). (C) Immunoblot analysis of GFP expression in cross-kingdom RNAi reporter plants at four time points upon B. cinerea infection using a @GFP antibody. The ribulose 1,5-bisphosphate carboxylase/oxygenase (RuBisCo) signal detected by Coomassie Brilliant Blue (CBB) staining was used as a protein loading control. Numbers indicate GFP and RuBisCo signal intensities estimated by the FIJI software. (D) Immunoblot analysis of SlAGO1 IP from leaf tissue infected with B. cinerea WT or Δbcago2 ko mutants using an anti-AGO1 antibody. RuBisCo signal detected by CBB staining was used as a protein loading control. (E) Stem-loop RT-PCR of BcsRNA3.1, BcsRNA3.2, and BcsRNA20 using SlAGO1 IP samples. Total RNA samples were used as a control for BcsRNA accumulation in B. cinerea WT and Δbcago2 ko mutant samples. The SlmiR399 was used as a control for the successful small RNA isolation from the SlAGO1 IP fraction, and BcTubulin and SlActin2 mRNA were used as a control indicating no unspecific RNA-binding. M: 10 bp DNA ladder.

Fig. 3.

BcAGO2 is a pathogenicity factor in B. cinerea. (A) Relative mRNA expression of BcAGOs in axenic culture or in Sl-infected samples at 48 hpi measured by qRT-PCR. B. cinerea BcTubulin was used as a housekeeping gene. (B) Tomato leaf infection and agar plate growth assays with B. cinerea WT and different Δbcago mutant strains. Leaf images were taken at 48 hpi to measure lesion area. Colony size on agar plates was measured in four independent cultures per strain at 3 d post cultivation start. (C) Relative B. cinerea genomic DNA (gDNA) was quantified in Sl-infected samples at 48 hpi by qRT-PCR, measuring tomato gDNA as a reference. In all plots, data points represent biological replicates and error bars indicate the SD. The letters indicate significant difference using one-way ANOVA and Tukey test with P < 0.05.

Fig. 4.

Small RNA binding profiles of BcAGO1 and BcAGO2 change during tomato infection. (A) Size distribution of total small RNA reads binding to BcAGOs in axenic culture with two biological replicates or in Sl-infected samples with three biological replicates at 48 hpi, as revealed by BcAGO IP and small RNA-seq. (B) Relative fractions of BcAGO-bound total BcsRNA reads mapping to different annotated genetic loci in the B. cinerea genome, including transfer RNA (tRNA), small nuclear and small nucleolar (sn/snoRNA), messenger RNA (mRNA), RT. (C) Normalized read counts per million (RPM) of different BcsRNAs binding to BcAGOs in axenic culture or Sl-infected samples. (D) Heatmaps of the size profile for BcAGO-bound BcsRNAs mapping to defined BcsRNA loci. Color scales indicate locus normalized expression, showing expression as a proportion of the highest-expressed size. Axenic culture and Sl-infected samples are shown on blue- and red-color scales, respectively.

Fig. 5.

Tomato small RNAs bind to BcAGOs during infection and induce cross-kingdom RNAi of B. cinerea genes. (A) Size distribution of SlsRNAs binding to BcAGO1 or BcAGO2 in Sl-infected samples, as revealed by BcAGO IP and small RNA-seq. (B) Left, B. cinerea was reisolated from Sl-infected material and grown in subculture on agar plates. Subculture samples were collected from the colony edge. Right, stem-loop RT-PCR of selected SlsRNAs detected in two biological replicates (#) of reisolated B. cinerea samples collected at 20 h or 7 d post cultivation start. Noninfected tomato leaves (Sl) were used as a positive control, and B. cinerea that was not reisolated from infected tomato as well as water were used as negative controls. The SlmiR399 was used as another negative control representing a highly expressed SlsRNA which was never detected in the BcAGO IP datasets. (C) Relative mRNA expression of predicted B. cinerea genes targeted by SlsRNAs was measured in Sl-infected samples at 48 hpi. B. cinerea BcActin was used as a housekeeping gene. (D) Tomato leaf infection and agar plate growth assays with B. cinerea WT and two independent target gene ΔBcin05g04730 ko strains (#). Leaf images were taken at 48 hpi to measure lesion area. Colony size on agar plates was measured in six independent cultures per strain at 3 to 5 d post cultivation start. In all plots, data points represent biological replicates. Error bars indicate the SD and numbers indicate significant difference using one-way ANOVA and the Tukey test with P < 0.05. (E) Alignments of the native Bcin05g04730 target site of SlsRNA12 or a target site scrambled version. (F) RNA in vitro cleavage assay using recombinant BcAGO that was preincubated with the SlsRNA12 and mixed with the native or scrambled Bcin05g04730 target RNA, as template. GFP was used as a nonspecific protein control. Template RNA was in the size of 94 bases and the expected cleavage products of 59 and 35 bases.