A Signaling Cascade from miR444 to RDR1 in Rice Antiviral RNA Silencing Pathway

Abstract

Plant RNA-DEPENDENT RNA POLYMERASE1 (RDR1) is a key component of the antiviral RNA-silencing pathway, contributing to the biogenesis of virus-derived small interfering RNAs. This enzyme also is responsible for producing virus-activated endogenous small interfering RNAs to stimulate the broad-spectrum antiviral activity through silencing host genes. The expression of RDR1 orthologs in various plants is usually induced by virus infection. However, the molecular mechanisms of activation of RDR1 expression in response to virus infection remain unknown. Here, we show that a monocot-specific microRNA, miR444, is a key factor in relaying the antiviral signaling from virus infection to OsRDR1 expression. The expression of miR444 is enhanced by infection with Rice stripe virus (RSV), and overexpression of miR444 improves rice (Oryza sativa) resistance against RSV infection accompanied by the up-regulation of OsRDR1 expression. We further show that three miR444 targets, the MIKC(C)-type MADS box proteins OsMADS23, OsMADS27a, and OsMADS57, form homodimers and heterodimers between them to repress the expression of OsRDR1 by directly binding to the CArG motifs of its promoter. Consequently, an increased level of miR444 diminishes the repressive roles of OsMADS23, OsMADS27a, and OsMADS57 on OsRDR1 transcription, thus activating the OsRDR1-dependent antiviral RNA-silencing pathway. We also show that overexpression of miR444-resistant OsMADS57 reduced OsRDR1 expression and rice resistance against RSV infection, and knockout of OsRDR1 reduced rice resistance against RSV infection. In conclusion, our results reveal a molecular cascade in the rice antiviral pathway in which miR444 and its MADS box targets directly control OsRDR1 transcription.

Figure 1.

miR444 accumulation was induced by virus infection. A, Analysis of the accumulation of miR444 and miR528 by RNA gel-blot hybridization in RSV- and mock-infected rice plants. Total RNA samples for hybridization were collected at 7 and 14 dpi. The 5S ribosomal RNA (rRNA) bands were visualized by ethidium bromide staining and served as a loading control. B, Quantitative reverse transcription (qRT)-PCR analysis of the expression of OsMADS23, OsMADS27a, and OsMADS57 in RSV-infected and mock control rice plants. After RT from the total RNA samples described in A (7 dpi), the relative mRNA levels were determined by qRT-PCR. Results are means ± se for three replicates. Asterisks indicate significant differences between infected and mock control rice plants. (Student’s t test analysis, P < 0.05). C, The accumulation of OsMADS23, OsMADS27a, and OsMADS57 proteins was analyzed by western blot at 7 dpi of RSV infection. The Rubisco large subunit (RBCL) bands were visualized by Coomassie Brilliant Blue and served as a loading control.

Figure 2.

Overexpression of miR444 improves rice resistance against RSV infection. A, Wild-type (WT) and miR444-overexpressed (2-1 and 13-1) rice plants were grown for 14 d and then inoculated by RSV. Protein samples were collected, and the accumulation of RSV CP was analyzed by western blot at 14 dpi of RSV infection. The Rubisco large subunit (RBCL) bands were visualized by Coomassie Brilliant Blue and served as a loading control. B, Growth performance of wild-type and miR444-overexpressed rice plants at 30 dpi of RSV infection.

Figure 3.

miR444 positively regulates the expression of OsRDR1 constitutively and upon virus infection. A, Overexpression of miR444 increased the expression of OsRDR1 constitutively. RNA samples were collected from wild-type (WT) and miR444-overexpressed rice plants (grown for 14 d). OsRDR1 mRNA levels relative to that of the wild-type rice plant were determined by qRT-PCR. B, Overexpression of miR444 markedly increased the expression of OsRDR1 during virus infection. Wild-type and miR444-overexpressed rice plants (grown for 14 d) were inoculated by RSV. At 14 dpi, RNA samples were collected, and OsRDR1 mRNA levels relative to that of the mock-inoculated wild-type rice plant were determined by qRT-PCR. Results are means ± se for three replicates. Asterisks indicate significant differences between the wild-type plant and miR444-overexpressed lines (Student’s t test analysis, P < 0.05).

Figure 4.

Dimerization of OsMADS23, OsMADS27a, and OsMADS57. A and B, Interaction patterns between OsMADS23, OsMADS27a, and OsMADS57 in Y2H assays. The yeast AH109 stain was cotransformed with the indicated constructs and grown on selective synthetic dropout (SD) medium. The construct was labeled by the gene name after AD (GAL activation domain) or BD (GAL-binding domain). 3AT, 3-Amino-1,2,4-triazole. C, Interaction patterns between OsMADS23, OsMADS27a, and OsMADS57 in GST pull-down assays. OsMADS23, OsMADS27a, and OsMADS57 were fused with GST and maltose-binding protein (MBP) tags and applied for GST pull-down assays. The interactions between OsMADS23, OsMADS27a, and OsMADS57 were detected by western blot using MBP antibody. Arrowheads indicate the expected bands of western blots. D, Interaction patterns between OsMADS23, OsMADS27a, and OsMADS57 in BiFC assays by cotransfecting rice protoplasts with the indicated constructs. Merged images of CFP fluorescence and rice protoplast are shown. The construct was labeled by the gene name followed by SCYNE (the N-terminal fragment of CFP) or SCYCE (the C-terminal fragment of CFP). The previously reported interacting proteins IPA1 (Ideal Plant Architecture1) and PCF2 (Proliferating cell nuclear antigen gene promoter binding factor 2) (Lu et al., 2013) were used as a positive control. The interactions between the three MADS proteins and IPA1 or PCF2 were detected and used as negative controls.

Figure 5.

OsMADS23, OsMADS27a, and OsMADS57 repress the expression of OsRDR1. A, Schematic structures of the effector and reporter constructs for the transient expression assay in N. benthamiana leaves, in which OsMADS23, OsMADS27a, and OsMADS57 were under the control of the Cauliflower mosaic virus (CaMV) 35S promoter, the GUS reporter gene harboring an intron was driven by the OsRDR1 promoter, and the Renilla luciferase (LUC; from Renilla reniformis) gene derived by the 35S promoter was used as an internal reference. B, Transient expression assay in N. benthamiana leaves. Relative GUS activities were normalized to the activities of Renilla luciferase and averaged from three biological repeats. Error bars indicate se. Asterisks indicate significant differences compared with the empty vector samples (Student’s t test analysis, P < 0.05).

Figure 6.

OsMADS23, OsMADS27a, and OsMADS57 bind to the CArG motifs of the OsRDR1 promoter. A, Schematic structures of the effector and reporter constructs for the Y1H assay. OsMADS23, OsMADS27a, and OsMADS57 were fused to the GAD, and the reporter gene LacZ was driven by the OsRDR1 promoter fragment containing a CArG motif. B, OsMADS23, OsMADS27a, and OsMADS57 bound to the promoter fragment of OsRDR1 in yeast. C to E, EMSAs showed that OsMADS23, OsMADS27a, and OsMADS57 bound to the CArG motif of the OsRDR1 promoter. The biotinylated probe containing the CArG motif sequence was incubated with GST-OsMADS23, GST-OsMADS27a, or GST-OsMADS57, while the probe incubated with no protein or GST protein was used as a negative control. Nonlabeled probes were used as cold competitors. F and G, ChIP-PCR assays showed that OsMADS23, OsMADS27a, and OsMADS57 bound to the OsRDR1 promoter regions containing CArG motifs (red boxes) in rice plants. Two pairs of PCR primers (P1/P2 and P3/P4) are indicated, and the amplified PCR bands are indicated by arrows. Asterisks indicate the dimers formed by primers. Ab, Antibody.

Figure 7.

Overexpression of miR444-resistant OsMADS57 (named OsMADS57^R) reduces OsRDR1 expression and rice resistance against RSV infection. A, Analysis of the mRNA levels of OsMADS57^R by qRT-PCR in transgenic rice lines (8-15 and 11-8) overexpressing miR444-resistant OsMADS57. B, Analysis of the accumulation of OsMADS57 protein by western blot in transgenic rice lines (8-15 and 11-8) overexpressing miR444-resistant OsMADS57. The arrowhead indicates the bands of OsMADS57. C, Analysis of the expression levels of OsRDR1 by qRT-PCR in OsMADS57^R-overexpressed rice plants relative to that of wild-type (WT) rice plants at 14 dpi of RSV infection. D and E, Analysis of the accumulation of RSV CP by western blot and the RSV CP RNA by qRT-PCR in OsMADS57^R-overexpressed and wild-type rice plants at 14 dpi of RSV infection. The Rubisco large subunit (RBCL) bands were visualized by Coomassie Brilliant Blue and served as a loading control in western blots. The qRT-PCR results are means ± se for three replicates. Asterisks indicate significant differences between wild-type and OsMADS57^R-overexpressed rice plants (Student’s t test analysis, P < 0.05).

Figure 8.

Knockout of OsRDR1 reduces rice resistance against RSV infection. A, Schematic structure of the Tos17 insertion site of the Osrdr1 mutant (ND0031). B, Homozygous Osrdr1 was verified by PCR assays using OsRDR1-specific primers (P1+P2) and OsRDR1- and Tos17-specific primers (P1+P3). C, RT-PCR analysis using gene-specific primers showing that the expression of OsRDR1 was knocked out in the Osrdr1 mutant. D, Analysis of the accumulation of RSV CP by western blot in Osrdr1 mutant and wild-type (WT) rice plants at 14 dpi of RSV infection. The Rubisco large subunit (RBCL) bands were visualized by Coomassie Brilliant Blue and served as a loading control. E, Analysis of the levels of the RSV CP RNA by qRT-PCR in the Osrdr1 mutant relative to that of wild-type rice plants at 14 dpi of RSV infection. Results are means ± se for three replicates. The asterisk indicates a significant difference between wild-type and Osrdr1 rice plants (Student’s t test analysis, P < 0.05).

Figure 9.

Model of the miR444-RDR1 signaling cascade in the rice antiviral RNA-silencing pathway. Three miR444 targets, OsMADS23, OsMADS27a, and OsMADS57, form homodimers, heterodimers, or even polymers between them to repress the expression of OsRDR1 by binding directly to the CArG motifs of its promoter. Upon virus infection, miR444 expression is induced. Consequently, an increased level of miR444 diminishes the repressive roles of OsMADS23, OsMADS27a, and OsMADS57 on OsRDR1 transcription. Then, the OsRDR1-dependent RNA-silencing pathway is activated to defend against viral infection by producing vsiRNAs to directly silence viral RNAs and virus-activated siRNAs (VasiRNAs) to silence host genes for the activation of broad-spectrum antiviral activity.