A Phase-Separated SR Protein Reprograms Host Pre-mRNA Splicing to Enhance Disease Susceptibility

Abstract

Alternative splicing (AS) plays a vital role in the plant-microbe interaction. Modulating host precursor-mRNA AS is a key strategy employed by multiple pathogens to subvert plant immunity. However, the underlying mechanism by which the host splicing factor regulates plant immunity remains poorly understood. Here, a plant-conserved serine/arginine-rich (SR) RNA splicing factor, SR30, which negatively regulates tomato immunity against the infamous Phytophthora infestans (P. infestans) is identified. SR30 governs tomato mRNA AS at a genome-wide level and suppresses defense-related genes AS. During P. infestans infection, SR30 is induced to form nuclear condensates via liquid-liquid phase separation driven by intrinsically disordered regions. Importantly, the phase separation property is required for the function of SR30 in disease susceptibility and the regulation of genes AS. Knockout of SR30 via CRISPR/Cas9 improves tomato disease resistance to P. infestans, P. capsici, and P. parasitica by promoting defense genes AS. These findings uncover a novel mechanism in a phase-separated protein that regulates plant immunity by altering the AS of defense-related genes and provides a new paradigm for engineering protein condensate in crop-resistant breeding.

Keywords: SR proteins; alternative splicing; late blight disease; phase separation; plant immunity.

Figure 1

SR30 negatively regulates tomato immunity. a) The heatmap showing the transcript levels of 18 genes encoding SR family proteins in tomato leaves in response to Phytophthora infestans infection. The data are derived from RNA‐seq in terms of FPKM (Fragments Per Kilobase of transcript per Million mapped reads) as reported previously.^{[ 30 ]} Asterisks represent the candidate SR family proteins selected for further immunity function analysis. b) Immuno‐blot analysis of GFP and six SR proteins expressed in Nicotiana benthamiana leaves using anti‐GFP antibody. Asterisks indicate the protein bands for each construct. Protein loading was visualized by Ponceau S staining. c) Relative infection lesion areas of P. infestans inoculated N. benthamiana leaves expressing SR proteins and GFP. The lesion areas were measured at 5 days post‐inoculation (dpi) and then normalized to the GFP control. Data represents the mean with standard errors (SE) (n = 12). P values were analyzed by Student's t‐test (**, P < 0.01). The original results of lesion areas for each SR protein are provided in Figure S1c. d) Schematic diagram of the construct used for generating SR30 transgenic lines. The coding sequence of tomato SR30 was fused with the N‐terminal GFP, which was driven by the CaMV 35S promoter and followed by a terminator. e) Relative expression of SR30 in two independent overexpressed (SR30‐OE) transgenic tomato (Micro‐Tom) lines was determined by quantitative reverse transcription‐polymerase chain reaction (qRT‐PCR). Data represents the mean with SE (n = 3). P values were analyzed by Student's t‐test (**, P < 0.01). The gene SlActin was used as an internal control. f) Protein detection of GFP‐SR30 in two SR30‐OE tomato (Micro‐Tom) lines by western blotting using anti‐GFP antibody. The protein bands of GFP‐SR30 are marked with asterisks. Protein loading was visualized by Ponceau S staining. g) Lesion area (cm²) of tomato (Micro‐Tom) leaves infected by P. infestans at 4 dpi. Data represents the mean with SE (n = 21). P values were analyzed by Student's t‐test (**, P < 0.01). h) Detection of the burst of reactive oxygen species in tomato (Micro‐Tom) leaves of WT and two SR30‐OE lines in response to 100 nm flg22. The leaf discs were equilibrated in sterilized water in a 96‐well plate for ≈8 h before treatment with flg22. i) Relative expression levels of three tomato defense‐related genes, LRR22, GRAS2, and WRKY28, were quantified in WT and two SR30‐OE transgenic tomato (Micro‐Tom) lines by qRT‐PCR. The gene SlActin was used as an internal control. The leaf samples were collected at 1 dpi. Data represents the mean with SE (n = 3). P‐values were analyzed by Student's t‐test (**, P < 0.01). j) Relative expression of SR30 in silencing tomato (Heinz 1706) plants. TRV2‐GFP was used as a negative control. The gene SlActin was used as an internal control. Data represents the mean with SE (n = 3). P values were analyzed by Student's t‐test (*, P < 0.05). k) Photographs of P. infestans infection assay on silenced tomato (Heinz 1706) leaves. P. infestans JH19 zoospores were inoculated on the detached tomato leaves 4 weeks after agroinfiltration. l) Lesion area (cm²) of P. infestans‐inoculated silenced tomato (Heinz 1706) leaves. Data represents the mean with SE (n = 36). The lesion areas were measured at 4–5 dpi. P values were analyzed by Student's t‐test (**, P < 0.01).

Figure 2

SR30 regulates AS of defense‐related genes in tomato. a) Schematic diagram of transcriptome sequencing using the Illumina platform. The tomato (Micro‐Tom) leaves of SR30‐OE and WT plants were sampled for RNA extraction and double‐stranded cDNA library synthesis. All RNA‐seq data were subjected to differential gene expression and differential alternative splicing analysis. b) The number (NO.) of differentially alternatively spliced events and corresponding genes identified from the comparison of SR30‐OE versus WT. c) NO. of differentially expressed genes (DEGs) were identified from the comparison of SR30‐OE versus WT. Up and down represent upregulated and downregulated DEGs, respectively. d) Venn diagram illustrating the number of differentially alternatively spliced genes (DASGs) and DEGs in the comparison of SR30‐OE versus WT. e–g) Validation of the splicing ratio of intron‐spliced and intron‐retained isoforms of three tomato defense‐related genes. The first panel from the left displays wiggle plots of RNA‐seq data for three selected genes, ubiquitin‐conjugating enzyme 34 (UBC34) (e), carbon catabolite repressor 4E (CCR4E) (f), and actin‐related protein 8 (ARP8) (g), with the schematic gene models of two different transcript isoforms. Asterisks indicate a premature termination codon (PTC). The second panel from the left shows the relative transcript level of two isoforms of three candidate genes using isoform‐specific primers by qRT‐PCR. The third panel shows the splicing ratio (transcript level of intron‐spliced isoform divided by transcript level of intron‐retained isoform) for three defense‐related genes. The tomato ubiquitin gene UBI was used as an internal control. The relative transcript level was normalized to the intron‐retained isoform of WT. Data represent the mean with SE (n = 3). P values were analyzed by Student's t‐test (** P < 0.01). The last panel shows the functions of different isoforms of three defense‐related genes in plant immunity against P. infestans. The GFP was used as a control. Lesion areas were measured at 5 or 6 dpi and then normalized to the GFP control. Data represents the mean with SE (n = 20). P‐values were analyzed by Student's t‐test (** P < 0.01).

Figure 3

SR30 forms nuclear condensates through LLPS. a) Subcellular localization of GFP and GFP‐SR30. GFP and GFP‐SR30 were transiently expressed in transgenic N. benthamiana leaves expressing a nuclear marker (H2B‐RFP) for 48 h before imaging by a confocal microscope (20× and 63× objective). Scale bars represent 10 µm. b) Subcellular localization of GFP‐SR30 in SR30‐OE transgenic tomato leaves. Scale bars represent 5 µm. c) FRAP of GFP‐SR30 expressed in epidermal cells of N. benthamiana leaves. The white arrows indicate the photobleached nuclear condensate. Scale bars represent 10 µm. d) Relative fluorescence recovery curves of photobleached GFP‐SR30. Data were normalized by the web‐based tool easyFRAP. Data represents the mean with standard deviation (SD) (n = 10). e) Fusion of two independent nuclear condensates of GFP‐SR30. The arrow indicates two nuclear condensates undergoing fusion. Scale bars represent 10 µm. f) Droplet formation of the purified protein GFP‐SR30 in vitro. The top cartoon shows the schematic diagram of His‐MBP‐GFP and His‐MBP‐GFP‐SR30 recombinant proteins. TEV protease was used to remove the His‐MBP tag. The bottom images show the droplet formation of 1 µM GFP‐SR30 in vitro in the presence of 150 mm NaCl. The GFP protein was used as a control. Scale bars represent 10 µm.

Figure 4

LLPS of SR30 is driven by IDRs. a) Predicted intrinsically disordered regions (IDRs) of SR30 using PONDR. Three gray boxes represent the region of RNA recognition motif (RRM), pseudo‐RNA recognition motif (ψRRM), and serine/arginine‐rich (SR) domain of SR30, respectively. b) Schematic diagrams of three designed IDR‐deletion mutants of SR30. c) Subcellular localization of IDR‐deletion variants of SR30 and complementary mutant SR30^ΔIDR2&5‐FUS. All mutants were expressed in nuclear marker H2B‐RFP transgenic N. benthamiana leaves for 48 h before imaging using a confocal microscope. Scale bars represent 10 µm. d) Subcellular localization of IDR‐deletion mutants of SR30 and SR30^ΔIDR2&5‐FUS under 1,6‐hexanediol (1,6‐Hex) treatment. All proteins were expressed in N. benthamiana for 36–48 h. For mock or 1,6‐Hex treatment, leaf tissues were infiltrated with Milli‐Q water or 5% 1,6‐Hex before observation. Scale bars represent 10 µm. e) Percentage of nuclei containing nuclear condensates of IDR‐deletion mutants and complementary mutant SR30^ΔIDR2&5‐FUS treated with 1,6‐Hex compared to mock control. A total of 300 nuclei were calculated for each treatment. f) Droplet formation of 1 µM GFP‐SR30 and GFP‐SR30^ΔIDR2&5 in vitro in the presence of 150 mm NaCl. Scale bars represent 10 µm.

Figure 5

SR30 regulates the AS of defense genes through LLPS to perform susceptible functions. a) Lesion area (cm²) of N. benthamiana leaves expressing GFP, GFP‐SR30, GFP‐SR30^ΔIDR2&5, and GFP‐SR30^ΔIDR2&5‐FUS at 5 dpi. Data represents the mean with SE (n = 16). P‐values were analyzed by Student's t‐test (**, P < 0.01). b) Schematic diagram of the procedure to validate the effect of SR30′s phase separation mutants on the AS of target genes. Step 1, the Agrobacterium tumefaciens strains carrying 35S::target genes gDNA (35S::SlUBC34/SlCCR4E/SlARP8 gDNA) were infiltrated into the whole leaves of N. benthamiana. Step 2, GFP, GFP‐SR30, GFP‐SR30^ΔIDR2&5, and GFP‐SR30^ΔIDR2&5‐FUS were expressed in different regions of the same N. benthamiana leaf expressing 35S:: target genes gDNA, respectively, at one‐day post‐infiltration (dpi). Step 3, the N. benthamiana leaves expressing different proteins were collected, respectively, and the total RNA of all treatment samples was extracted to perform qRT‐PCR assays. c) The splicing ratio of two different isoforms of UBC34, CCR4E, and ARP8 under treatment with GFP, GFP‐SR30, GFP‐SR30^ΔIDR2&5, and GFP‐SR30^ΔIDR2&5‐FUS by qRT‐PCR, respectively. The gene NbActin was used as an internal control. For each target gene, the relative transcript level of two different isoforms of different treatments is normalized to the transcript level of the intron‐spliced isoform. Data represent the mean with SE (n = 6). P values were analyzed by Student's t‐test (**, P < 0.01). This experiment was repeated twice with similar results. d) Schematic diagram of the split‐LUC assays to test the interactions between SR30 and other splicing factors (SFs). SR30, SR30^ΔIDR2&5, and SR30^ΔIDR2&5‐FUS were inserted into the pICH86988‐nLUC vector, while other splicing factors were inserted into the pCAMBIA1300‐cLUC vector. If two test proteins interact with each other, the luciferase will emit fluorescence in the presence of luciferin. e) Relative LUC activity of the interaction between SR30, SR30^ΔIDR2&5, and SR30^ΔIDR2&5‐FUS with other tomato splicing factors by split‐LUC assays, respectively. Leaf discs were used to measure the luminescence 48 h after co‐expression of the indicated proteins. The relative LUC activity of SR30^ΔIDR2&5 and GFP‐SR30^ΔIDR2&5‐FUS with each splicing factor was normalized to SR30 with the same splicing factor. Data represents the mean with SE (n = 13). P‐values were analyzed by Student's t‐test (*P < 0.05, **P < 0.01).

Figure 6

Knockout of SR30 enhances tomato resistance against three oomycetes pathogens. a) Formation of nuclear condensates of GFP‐SR30 driven by the native SR30 promoter was examined under P. infestans infection. The agrobacterium carrying proSR30::GFP‐SR30 was infiltrated into N. benthamiana leaves. The infiltrated areas of leaves were inoculated with either mock (Milli‐Q water) treatment or P. infestans zoospores at one‐day post‐agroinfiltration. Leaf tissues were sampled at zero‐ and three‐days post‐inoculation of mock or P. infestans, observed under a confocal microscope, and photographed using ZEN software. Scale bars represent 10 µm. b) Schematic diagram of two small guide RNAs (sgRNAs) designed to knockout SR30. Arrows represent two designed sgRNA positions in the SR30 gene. Red boxes represent exons and black lines represent introns. c) Genomic DNA (gDNA) sequence of SR30 in two different sr30 mutants. The red nucleobases indicate the positions of two protospacer adjacent motifs (PAM). The red regions indicate the differences of SR30 gDNA sequences in two sr30 mutants compared to WT by sequence alignment. d) Protein sequences of SR30 in sr30 mutants and WT tomato. The asterisk indicates the stop codon. e–g) The image showed that knockout SR30 in tomatoes suppressed the lesion growth of P. infestans (e), P. capsici (f), and P. parasitica (g). For P. infestans, detached leaves of sr30 and WT were inoculated with zoospores. Infected leaves were stained with trypan blue and photographed 4 dpi. For P. capsici and P. parasitica, detached leaves of sr30 and WT were inoculated with mycelium disc. Infected leaves were stained with trypan blue and photographed 2 dpi. The columns showed the lesion areas of tomato (Micro‐Tom) leaves infected by three Phytophthora pathogens. For (e–g), data represent the mean with SE (n = 17). P values were analyzed by student′s t‐test (*P < 0.05, **P < 0.01).

Figure 7

The knockout of SR30 promotes the AS of defense‐related genes during pathogen infection. a,b) Splicing ratio of UBC34, CCR4E, and ARP8 in sr30 tomato mutants during the non‐infected condition (a) and P. infestans infection (b). The infected leaves samples were collected at 2 dpi with P. infestans. The tomato ubiquitin gene UBI was used as an internal control. Data represent the mean with SE (n = 6). P‐values were analyzed by Student's t‐test (*P < 0.05, **P < 0.01). For (a,b), each experiment was repeated twice with similar results. c) A proposal working model of SR30 negatively regulating tomato immunity. During P. infestans infection, the increased concentration of SR30 promotes the formation of nuclear condensates via LLPS where SR30 sequestrates other splicing factors and compromises spliceosome function in WT. As a result, the AS of defense‐related genes is suppressed and the plant becomes more susceptible. Conversely, the truncated SR30 cannot form condensates to destabilize defense genes AS in the sr30 mutant. Consequently, the loss of function of SR30 confers tomato resistance to P. infestans.