Rhomboid-mediated cleavage of the immune receptor XA21 protects grain set and male fertility in rice

Abstract

To maintain growth and to successfully reproduce, organisms must protect key functions in specific tissues, particularly when countering pathogen invasion using internal defensive proteins that may disrupt their own developmental processes. The rice immune receptor XA21 confers race-specific resistance against Xanthomonas oryzae pv. oryzae, which causes the deadly disease bacterial leaf blight. Here, we demonstrate that XA21 is cleaved by the rhomboid-like protease OsRBL3b, likely within its transmembrane domain. OsRBL3b mRNA transcripts are preferentially expressed in rice spikelets. Rice plants expressing Xa21 but lacking a functional OsRBL3b displayed impaired anther dehiscence and pollen viability, resulting in male sterility and yield reduction with high levels of XA21 protein present in spikelets during anthesis. In leaves, osrbl3b mutants expressing XA21 had normal levels of this resistance protein and disease immunity. This balance between reproduction and disease resistance through the specific expression of a rhomboid protease may be key to limiting the detrimental effects of an active immune response and may be useful in future for genetic improvement of crops.

Keywords: Oryza sativa; disease resistance; innate immune receptor; intramembrane protease.

Fig. 1.

Expression of OsRBL3b and accumulation of XA21 in rice spikelets at anthesis. (A) OsRBL3b is preferentially expressed in spikelets. Levels of mRNA were quantified by RT-qPCR using the Myc-Xa21^L line. Results were normalized to the expression of the housekeeping genes UBIQUITIN5 and ACTIN. Expression levels in the indicated tissues were expressed relative to those of 2-mo-old leaves, which were arbitrarily set to 1 unit. Values are means ± SD of three biological replicates, each with three technical replicates. DAF, days after flowering. BA, before anthesis. (B) Accumulation of Myc-XA21 protein in spikelets. Myc-XA21 protein levels were determined in the indicated lines by immunoblot using an anti-Myc antibody, with ATPase levels as loading controls. WT^sib lacking XA21 is a negative control. Spikelets from more than three plants per line were used for protein extraction. (C) Quantification of Myc-XA21 from (B). This immunoblotting was repeated three times with similar results (all replicates in this study are shown in SI Appendix). In each experiment, the Myc-XA21 protein level was normalized against the ATPase level. The Myc-XA21 protein level in Myc-Xa21^L was set to 1.0 as references for calculating the relative level of Myc-XA21 in Myc-Xa21^H. Statistical analyses were performed using Student’s t-test. Asterisks denote statistically significant differences (*P < 0.05, **P < 0.01).

Fig. 2.

OsRBL3b mediates cleavage of XA21 in N. benthamiana. (A) Diagram of the Xa21 construct used for the cleavage assay in N. benthamiana. LRR, leucine-rich repeat domain; TMD, transmembrane domain. c-Myc and 3xFLAG epitope tags were fused to XA21 for detecting the N terminus and C-terminus of the protein, respectively. The intracellular kinase domain (IKD) used to generate the α-XA21IKD antibody is indicated by the black horizontal line. (B) Cleavage of XA21 depends on active OsRBL3b. Constructs encoding the indicated proteins were coexpressed in N. benthamiana. OsRBL3b^S187A is mutated in the catalytically essential residue Ser187. Total protein extracts were subjected to immunoblot analysis using an anti-Myc antibody, anti-XA21IKD, and anti-HA antibody, respectively. EV, empty vector. Asterisks indicate nonspecific cleavage products. Ponceau S staining of Rubisco large subunit was used as loading control. Ncp, N-terminally cleaved product; ccp, C-terminally cleaved product. (C) OsRBL3b specifically mediates XA21 cleavage. OsRBL3a is closely related to OsRBL3b (SI Appendix, Fig. S2). Protein samples in B and C were prepared independently.

Fig. 3.

OsRBL3b specifically controls the steady-state levels of XA21 in rice spikelets at anthesis. (A) Accumulation of Myc-XA21 protein in the spikelet of WT^sib, Myc-Xa21^L, and the osrbl3b Myc-Xa21^L mutants, determined by immunoblot analysis using an anti-Myc antibody and anti-ATPase as a loading control. Spikelets used for the experiment were collected from fresh plants. (B) Quantification of Myc-XA21 protein from A. (C) Relative Myc-Xa21 transcript levels, quantified by RT-qPCR, in spikelets of the lines shown in A. (D) Accumulation of Myc-XA21 protein in the spikelet of the osrbl3b-b Myc-Xa21^H and the Myc-Xa21^H lines. Spikelets used for the experiment were harvested from plants maintained by continuous ratooning. (E) Quantification of Myc-XA21 protein from D. (F) Relative Myc-Xa21 transcript levels, quantified by RT-qPCR, in spikelets of the osrbl3b-b Myc-Xa21^H and Myc-Xa21^H lines. (G) Myc-XA21 protein accumulation in leaves of WT^sib, Myc-Xa21^L, and osrbl3b Myc-Xa21^L. (H) Quantification of Myc-XA21 protein from G. Myc-XA21 protein levels in B, E, and H were first normalized with ATPase levels. For each experiment, the Myc-XA21 protein level in Myc-Xa21^L was set to 1.0 as references for calculating the relative level of protein Myc-XA21 in the other lines. Results in C and F were normalized relative to levels of mRNAs for UBIQUITIN 5 and ACTIN. Values are means ± SD of three biological replicates, each with three technical replicates. The immunoblot analyses were repeated three times with similar results. WT^sib (lacking Xa21) was used as a negative control for Myc-XA21 protein and Myc-Xa21 transcript abundance. Values in B, E, and H are the means ± SD (n = 3, biological replicates). Spikelets from more than three plants per line were used for protein or RNA extraction. Statistical analyses were performed using Student’s t-test. Asterisks indicate statistically significant differences (*P < 0.05, **P < 0.01); n.s., not statistically significant difference.

Fig. 4.

Mutation of OsRBL3b decreases grain-set but retains resistance in an XA21-dependent manner. (A) Panicle phenotypes of the Myc-Xa21^H line and the sterile osrbl3b-b Myc-Xa21^H mutant line. (B) Phenotypes of mature plants (Upper) and floral structure (Lower) of the indicated lines in A. (Scale bars, 1 mm.) Materials used for the experiments described in A and B were harvested from plants maintained by continuous ratooning in a greenhouse. (C) Panicle phenotypes of control lines and osrbl3b Myc-Xa21^L mutants grown in a temperature-controlled growth chamber at low temperature, showing lower fertility. (Scale bar, 1 cm.) (D) Phenotypes of mature plants (Upper) and florets (Lower) of the indicated lines in C. (Scale bar, 1 mm.) (E) Grain-setting rates of osrbl3b Myc-Xa21^L mutants and their controls. Values are shown as the mean ± SD from five independent plants with two sampled panicles each. Statistical analyses were performed using the Tukey–Kramer honestly significant difference test. Different letters indicate statistically significant differences (P < 0.05). This experiment was repeated two times in different seasons with similar results. (F) XA21 confers normal resistance in the osrbl3b mutant background. Indicated lines were inoculated with the Xoo strain PXO99^A by the leaf-clipping method. Inoculated leaves from the indicated lines showing lesion development at 12 days post inoculation (dpi). Arrows indicate the inoculation sites. (G) Lesion length was scored at 12 dpi. Each data point represents 17 inoculated leaves. Values are means ± SD. Statistical analyses were performed using the Tukey–Kramer honestly significant difference test. Different letters indicate statistically significant differences (P < 0.05). This experiment was repeated three times with similar results. (H) Rice seedlings carrying Xa21 survive Xoo infection regardless of the presence of a functional OsRBL3b. The photograph was taken 5 wk post inoculation. Arrows indicate the inoculation sites. Similar results were obtained from 10 inoculated individuals

Fig. 5.

Mutation of OsRBL3b impairs pollen viability and anther dehiscence in an XA21-dependent manner. (A) osrbl3b-b Myc-Xa21^H mutants display lower starch accumulation in their pollen grains. I₂⁻KI-stained pollen grains from Myc-Xa21^H and osrbl3b-b Myc-Xa21^H lines. Pollen viability was quantified based on the number of stained pollen grains (dark color) relative to the total pollen counted. Values are means ± SD, n = 3 (three independent samplings). Statistical analysis was performed using Student’s t-test. Asterisks indicate a statistically significant difference (**P < 0.01). This experiment was repeated two times with similar results. (B) osrbl3b Myc-Xa21^L mutants display lower starch accumulation in their pollen grains. (Left) I₂⁻KI-stained pollen grains from osrbl3b Myc-Xa21^L and the indicated control lines grown in a lower-temperature growth chamber before anthesis. (Right) Pollen quantified as above. Values are means ± SD (n = 3). Statistical analysis was performed using the Tukey–Kramer honestly significant difference test. Different letters indicate statistically significant differences (P < 0.05). (C) (Top 2 rows) I₂⁻KI staining of pollen grains in anthers of the indicated lines before and after anthesis. (Third row) Pollen grains on the stigmas of the indicated lines after anthesis. This experiment was repeated two times with similar results. (Bottom row) Cross-sections of anthers from the indicated lines after anthesis. Scale bars in A and B indicate 500 μm (rows 1, 2, and 4) or 50 μm (row 3). Materials used for the experiments described in this figure were harvested from plants maintained by continuous ratooning in a greenhouse. (D) osrbl3b mutations impaired anther dehiscence in the osrbl3b Myc-Xa21^L line. The graph shows the distribution of dehiscent and indehiscent spikelets among more than 100 random flower samples chosen from five individual plants of each indicated line. (E) Morphology of dissected spikelets from the indicated lines. (Top 2 rows) I₂⁻KI staining of pollen grains in anthers of the indicated lines before anthesis and 2 h after anthesis. (Third row) Presence or absence of pollen grains on the stigmas of the indicated lines after anthesis. Both dehisced and indehiscent anthers were shown. This experiment was repeated two times with similar results. (Bottom row) Cross-sections of anthers from the indicated lines after anthesis. [Scale bars, 500 μm (rows 1, 2, and 4) or 50 μm (row 3).]

Fig. 6.

Down-regulation of JA-responsive and JA-signaling genes and upregulation of defense-related genes in the indehiscent osrbl3b-b Myc-Xa21^H spikelets relative to the dehiscent Myc-Xa21^H spikelets at anthesis as detected by RNA-seq analysis. (A) RNA-seq volcano plot showing downregulation of a set of JA-responsive and JA-signaling genes. (B) Relative transcript levels of the JA-responsive and JA-signaling genes in A with RPKM (reads per kilo base per million mapped reads) >1, quantified by RT-qPCR, in spikelets of lines Myc-Xa21^H and osrbl3b-b Myc-Xa21^H. Results were normalized relative to levels of the mRNAs for UBIQUITIN 5 and ACTIN. Values are means ± SD of three biological replicates, each with three technical replicates. Statistical analyses were performed using Student’s t-test. Asterisks denote statistically significant differences (**P < 0.01). (C) RNA-seq volcano plot showing differentially expressed rice NLR genes. A complete list of transcript IDs detected is shown in the raw data file (Raw data for Fig. 6C). Cut-off criteria for data in A and C: log₂[osrbl3b Myc-Xa21^H/ Myc-Xa21^H] ≥ 1 and ≤ −1 and FDR-value <0.05.