Fig 1.
Phytophthora capsici undergoes dramatic changes in transcription and RNPII Ser5 phosphorylation during sporangia formation.
(A) Images of P. capsici during sporangia formation. The left panel shows the hypha growth state (HY) of P. capsici. The image was taken 2 days after P. capsici was inoculated on V8 medium and incubated at 25°C in the dark. The right panel shows the sporangia formation state (sporangia were produced at the tips of the hyphae; HYSP) of P. capsici. The image was taken 3 days post inoculation under the same conditions. Scale bar: 200 μm. (B) SDS-PAGE analysis of the proteomic profiles of P. capsici before (HY) or after (HYSP) sporangia development. (C, D, and G) Transcriptomic changes in P. capsici during sporangia formation. D shows the gene profile with altered expression levels during P. capsici sporangia formation, while C and G show the enriched KEGG and GO pathways of the altered genes. (E and F) RNPII undergoes an elevation of Ser5 phosphorylation in the CTD during sporangia formation. The relative intensity of Ser5-phosphorylated RNPII and unphosphorylated RNPII in the hypha stage (HY) and sporulated hypha stage (HYSP) was quantified with ImageJ (E). Data presented in F are as the mean ± SD of three replicates. Statistical significance compared to the HY was determined using Student’s t-test (* P < 0.01). β-tubulin was used as the loading control and its intensity in each sample was used to normalize the data between samples.
Fig 2.
PcDCL1 is crucial for sporangia development in Phytophthora capsici through precisely phosphorylating certain proteins including RNPII.
(A) Phylogenetic analysis of DCLs from different species. The phylogenetic tree was constructed by the neighbor-joining method. (B) The relative gene expression level of PcDCL1 at different life stages and infective stages. HY: hypha; HYSP: sporangia-generated hypha; SP: sporangia only; ZO: zoospores; CY: cysts; IN-0H/6H/12H/24H/72H: P. capsici inoculated onto chili leaves for 0, 6, 12, 24 or 72 h. The expression level in the HY stage was set as 1.0 and relative gene expression level was calculated by the 2-ΔΔCt method. The housekeeping gene WS21 was used as an internal control. (C) Sporangia development was severely impaired in PcDCL1 knockout mutants. The number of sporangia was measured and compared in wild-type P. capsici LT1534 (WT), empty vector isolate (EV), three PcDCL1 knockout mutants (koPcDCL1-1/2/3), and a PcDCL1 complementary isolate (comPcDCL1). Due to the relatively slower growth rate of PcDCL1 knockout mutants (on average about 50% slower in koPcDCL1 than other isolates), for koPcDCL1-1/2/3, the sporangia measurement was performed 7 days after inoculation on V8 medium; for other isolates, the same experiment was performed 3.5 d after inoculation. Data are presented as the mean ± SD of 10 biological replicates. Statistical significance compared to the WT was determined using Student’s t-test (* P < 0.05). (D) Gene ontology analysis of the genes differently expressed in WT and koPcDCL1. The analysis was based on the molecular function of the genes. (E) Phosphoproteomics analysis showed that the representative proteins with the largest differences in phosphorylation levels between WT and koPcDCL1 were mainly involved in RNA processing and DNA binding. The phosphorylated proteins from the sporangia-generated hypha stage of WT and koPcDCL1 were extracted and detected by mass spectrometry. (F and G) RNPII Ser5 phosphorylation was elevated in koPcDCL1 than WT and comPcDCL1. The relative intensity of Ser5-phosphorylated RNPII and unphosphorylated RNPII in the sporulated hypha stage of WT, koPcDCL1, and comPcDCL1 was quantified with ImageJ (F). The WT band detected by each antibody was given a value of 1.00. Data presented in G are as the mean ± SD of three replicates. Statistical significance compared to the WT was determined using Student’s t-test (* P < 0.01). β-tubulin was used as the loading control and its intensity in each sample was used to normalize the data between samples. (H) The morphology of wild-type LT1534 and koPcDCL1 in sporulated hypha stage (HYSP). LT1534 was grown for 3.5 days and koPcDCL1 was grown for 7 days on V8 medium under light, respectively. Almost no canonical sporangium but a plenty of abnormal hyphal branches (ab; indicated by black triangles) that might be the precursors of sporangiophores were observed only in koPcDCL1 mutants.
Fig 3.
pcamiR1 is a novel DCL1-dependent microRNA that plays an important role in sporangia development in P. capsici.
(A) Small RNA-seq revealed that four novel microRNAs were differentially expressed between the hypha stage (HY) and sporangia stage (SP). By secondary structure prediction of the DNA sequence from which small RNAs originated, pcamiR1 was identified as the only microRNA specifically and highly expressed in the sporangia stage. (B) The secondary structure of pcamiR1 precursor was predicted by Mirdeep and Mireap. (C) Relative gene expression level of pcamiR1 and its precursor in the hypha (HY) stage or sporulated hypha (HYSP) stage of wild-type P. capsici LT1534 (WT) and PcDCL1 knockout mutant koPcDCL1. For each gene, the expression level of WT in the HY stage was given a value of 1.0 and relative gene expression level was calculated by the 2-ΔΔCt method. The housekeeping genes WS21 and 5S rRNA were used as internal controls to quantify the pcamiR1 precursor and pcamiR1, respectively. (D) Relative gene expression level of pcamiR1 in the HYSP stage of WT, empty vector isolate (EV), kopcamiR1 mutants, and pcamiR1 complementary isolate compcamiR1. The expression level in WT was given a value of 1.0 and relative gene expression level was calculated by the 2-ΔΔCt method. The housekeeping gene 5S rRNA was used as an internal control. (E) Sporangia development was impaired in kopcamiR1 mutants. The number of sporangia in WT, EV, three pcamiR1 knockout mutants (kopcamiR1-1/2/3) and compcamiR1 was measured and compared. Due to the relatively slower growth rate of kopcamiR1 (on average about 30% slower in kopcamiR1 than other isolates), for kopcamiR1-1/2/3, the sporangia measurement was performed 6 days after inoculation onto V8 medium; for other isolates, the same experiment was performed 4 days after inoculation. Data in C, D, and E are presented as the mean ± SD of 3 (C and D) or 10 (E) biological replicates. Statistical significance compared to the WT or koPcdcl1 was determined using Student’s t-test (* P < 0.05). (F and G) RNPII Ser5 phosphorylation was significantly elevated in kopcamiR1 compared to WT and compcamiR1. Data presented in G are as the mean ± SD of three replicates. Statistical significance compared to the WT was determined using Student’s t-test (* P < 0.01). β-tubulin was used as the loading control and its intensity in each sample was used to normalize the data between samples. (H) Phenotypes of kopcamiR1. The upper lane shows the hypha growth status of WT, kopcamiR1, and compcamiR1; the lower lane shows the sporulated hypha growth status of the same isolates. The images were taken 6 days after kopcamiR1 was inoculated and 4 days after WT or compcamiR1 were inoculated.
Fig 4.
pcamiR1 could efficiently repress the expression of PcCDK7 through ribosome occupancy associated translational inhibition.
(A) Illustration of the predicted binding site between pcamiR1 and the 3’-UTR of PcCDK7 or its mutated sequence (ckts). (B) The stringently positive correlation of the gene expression levels between PcCDK7 and the pcamiR1 precursor at different life stages and infective stages. HY: hypha; HYSP: sporulated hypha; SP: sporangia only; ZO: zoospores; CY: cysts; IN-0H/6H/12H/24H/72H: P. capsici inoculated onto chili leaves for 0, 6, 12, 24 or 72 h. The expression level in the HY stage was given a value of 1.0 and relative gene expression level was calculated by the 2-ΔΔCt method. The housekeeping gene WS21 was used as an internal control. (C) Gene expression was not influenced in the pcamiR1 knockout mutant (kopcamiR1) compared to wild-type P. capsici LT1534 (WT). Statistical significance compared to the WT was determined using Student’s t-test (* P < 0.05). (D, E and F) The GFP tandem reporter system showed that the GFP signal was sharply weakened when the pcamiR1 target sequence of PcCDK7 (CDK7ts) and pcamiR1 were co-expressed. Both GFP fluorescence (D and E) and GFP protein density (F) were assessed; 35S promoter-driven GFP was linked with CDK7ts or a target region mutated CDK7ts sequence (ckts), and pcamiR1 was expressed using a frequently used plant artificial miRNA expression vector (pEG-amiR171). The GFP signal was detected by confocal microscopy 2 days after Nicotiana benthamiana infiltration, the GFP intensity of five images of each treatment were calculated by ImageJ. Data are presented as the mean ± SD. Statistical significance compared to each control were determined using Student’s t-test (* P < 0.01). For western blotting, proteins from leaves were extracted and the house keeping protein actin was used as the loading control, its intensity in each sample was used to normalize the data between samples. The band intensity was calculated by ImageJ. Scale bar: 100 μm. (G) In vivo translational inhibition assay of pcamiR1 on PcCDK7. PcamiR1 was overexpressed in FLAG-tagged PcCDK7 overexpression isolate, the translated PcCDK7 protein was detected by western blotting. P. capsici house keeping protein β-tubulin was used as the loading control, its intensity in each sample was used to normalize the data between samples. The band intensity was calculated by ImageJ. (H) Ribosome occupancy is the key to result in the translational inhibition of pcamiR1 on PcCDK7. Relative abundance of PcCDK7-mRNA in total RNA, none-ribosome fraction, 40-80S ribosome fraction, or polysome fraction was calculated by the 2-ΔΔCt method by comparing its abundance with the amount of the housekeeping gene WS21 in corresponding fractions. One of the WT samples in each fraction was set as 1.0 and statistical significance was determined using Student’s t-test (* P < 0.05). (I and K) pcamiR1 could efficiently repress the translation of PcCDK7 by binding to its 3’-UTR. The in vitro gene silencing assay was performed based on WGE. pcamiR1 and a control miRNA (miR8788), which is the only miRNA identified in Phytophthora so far (confirmed in P. capsici by small RNA-seq), PcAGO1 and the control protein PcAGO4, and GFP-CDK7ts and the control target site GFP-ckts were separately paired and equally incubated in the system. The GFP abundance was detected by western blotting. To further validate the translation inhibition effect, the RISC (pcamiR1 + PcAGO1) was added to an equal amount of GFP-CDK7ts DNA at different concentrations, and a positive relationship was found between the RISC concentration and gene silencing efficiency. The protein loading control is shown by Ponceau staining. (J) pcamiR1 preferentially bound to the argonaute protein PcAGO1. All the four canonical AGOs in P. capsici were labeled with 3×FLAG and expressed using the wheat germ cell-free expression system (WGE). pcamiR1 mimics were incubated with equal amounts of the four AGOs and co-immunoprecipitated after 3×FLAG bead enrichment. The abundance of pcamiR1 binding to each AGO protein was detected by northern blotting.
Fig 5.
PcCDK7 directly phosphorylates RNPII and modulates sporangia formation in P. capsici.
(A and B) PcCDK7 was localized in the nucleus. The subcellular localization of PcCDK7 was examined by confocal observation (A) and western blotting based on nuclear-/cytoplasmic-fraction isolation (B). Histone H3 and GAPDH were used as markers of nucleus and cytoplasm, respectively. (C) CDK7 inhibitor THZ1 had no effect on hypha growth in wild-type P. capsici LT1534 (WT) and the PcCDK7 overexpression isolate oePcCDK7. CK represents the DMSO treatment. (D and I) THZ1 could efficiently inhibit sporangia formation in WT but not in oePcCDK7. Sporangia development was impaired in WT grown on V8 medium with THZ1 added, while oePcCDK7 was not influenced by THZ1. Measurement and imaging were performed 4 days after the isolate was inoculated onto the medium. Data are presented as the mean ± SD of 10 biological replicates. Statistical significance compared to water treatment was determined using Student’s t-test (* P < 0.05). Scale bar: 100 μm. (E and F) RNPII Ser5 phosphorylation was elevated in oePcCDK7 than in WT and empty vector isolate (EV). The relative intensity of Ser5-phosphorylated RNPII and unphosphorylated RNPII in the sporangia-generated hypha stage of WT, EV, and oePcCDK7 was quantified with ImageJ. The WT band detected by each antibody was given a value of 1.00. Data presented in F are as the mean ± SD of three replicates. Statistical significance compared to the WT was determined using Student’s t-test (* P < 0.01). β-tubulin was used as the loading control and its intensity in each sample was used to normalize the data between samples. (G and H) THZ1 could repress RNPII Ser5 phosphorylation in P. capsici. LT1534 was grown on V8 medium with or without THZ1 for 5 days, the sporulated hypha was collected and used for protein extraction, and the relative intensity of Ser5-phosphorylated RNPII and unphosphorylated RNPII was quantified by western blotting. The WT-CK band detected by each antibody was given a value of 1.00. Data presented in H are as the mean ± SD of three replicates. Statistical significance compared to the WT-CK was determined using Student’s t-test (* P < 0.01). The protein loading control is shown by Ponceau staining.
Fig 6.
PcCDK7 interacts with the characteristic RNPII CTD and upregulates Ser5 phosphorylation in Phytophthora.
(A) Phylogenetic analysis of the largest subunit of RNPII from different species. The RNPII CTD is located in the C-terminal of the largest subunit of RNPII, which harbors heptapeptide repeats in most higher organisms and Phytophthora species. The phylogenetic tree was constructed based on the amino acid sequences of the largest subunit of RNPII from different species with MEGA 5.0 using the neighbor-joining method. (B) Structural comparison of RNPII CTD from Homo sapiens (hsaCTD) and Phytophthora capsici (pcaCTD). About 400 amino acids of RNPII CTD from the two species were extracted and modeled by AlphaFold. HsaCTD has a loose and liner structure, while pcaCTD is more compact and could form canonical secondary structures in its central region. (C) Sketch map of the canonical Phytophthora RNPII CTD and four mutated RNPII CTDs. The red letters indicate mutated amino acids compared to the wild-type CTD. (D) The interaction between PcCDK7 and RNPII was verified via in vivo Co-IP assay. Total proteins for each sample were extracted from each isolate expressing 3×FLAG-PcCDK7 or 3×FLAG-GFP. The immune complexes were pulled down using anti-FLAG magnetic beads and immunoblotted with an anti-FLAG and anti-RNPII Ser5p antibody. 3×FLAG-GFP was used as the negative control. (E) PcCDK7 interacted with RNPII CTD regardless of amino acid substitutions or lack of the only canonical heptapeptide YSPTSPS. 3×FLAG-PcCDK7 and GST-CTD or the mutated GST-CTDs were expressed by Escherichia coli isolate BL21. The in vitro pull-down assay was performed using GST-sepharose beads after incubating PcCDK7 with GST-labeled proteins or GST. Anti-FLAG and anti-GST Ser5p antibodies were used to detect the interaction. GST was used as the negative control. (F and G) PcCDK7 could efficiently phosphorylate RNPII CTD unless the Ser5 in the heptapeptide repeats was mutated. A pan-phosphorylated-Ser antibody was used to detect the phosphorylation states of the native RNPII CTD and mutated CTDs after incubation with PcCDK7 (upper panel). Data presented in G as the mean ± SD of three replicates. Statistical significance compared to the GST-CTD was determined using Student’s t-test (* P < 0.01). Each protein sample was normalized by immunoblotting with the anti-GST antibody as the loading control (lower panel).
Fig 7.
The phosphorylation state of RNPII determines its chromosomal occupation and induces different transcriptomes.
(A) The phosphorylation state of RNPII determines its chromosomal occupation. Immunoprecipitated DNA from hypo-/moderately/hyper-phosphorylated RNPII was sequenced and identified. The distribution of the identified peaks in different genome regions was assessed by BWA and MACS. (B) Similarity analysis of the three levels of phosphorylated RNPII’s binding profiles. (C) Identification of the binding motifs of RNPII with different phosphorylation levels using the multiple EM for motif elicitation (MEME) program. The top two motifs for each phosphorylation level are shown. (D) Venn diagram showing the peak distributions in moderately phosphorylated RNPII vs hypo-phosphorylated RNPII or hyper-phosphorylated RNPII. The analysis was performed by MAnorm, a region of overlap ≥ 50% of the peak was considered the common peak, and peaks with |M| ≥ 1 and −log10p-value ≥ 3 in the comparison were considered as differential peaks. (E) Venn diagram showing the number of genes both differentially expressed and differentially bound by RNPII in the two compared groups. Differentially expressed genes were defined as |log2fc| > 1 and FDR < 0.05 in an RNA-seq comparison. (F) The number of peak-related genes differentially bound by RNPII in the two compared groups. Peak-related genes were confirmed by their genomic location and gene annotation, the binding abundance was calculated based on the peak distribution in each gene. (G and H) Pathway enrichment analysis of genes differentially bound by RNPII in the two compared groups. The analysis was based on gene ontology analysis. (I and J) Four-quadrant diagram analysis showed the genes with the most significant changes in both RNA-seq and ChIP-seq. The labeled genes were uniformly variable in two sequencing datasets. Genes with log2fc > 1 and M > 0.5 (moderate/hypo) or 1 (hyper/moderate) are labeled with green numbers and genes with log2fc < −1 and M < −0.5 (moderate/hypo) or −1 (hyper/moderate) are labeled with blue numbers in the two compared groups. (K) Representative genes with the most significant changes in both RNA-seq and ChIP-seq in the two compared groups and (L) PcCDK7 and pcamiR1 precursor gene in hypha (hypo-phosRNPII) and sporulated hypha (moder-phosRNPII) stages. ChIP-seq data (dark blue) and RNA-seq data (light blue) were depicted in the Integrative Genomics Viewer (IGV) browser. Black boxes represent the open reading frames of the genes.
Fig 8.
Application of pcamiR1 antagomir could be a promising strategy for Phytophthora capsici control.
(A, B, C and E) Hypha growth rate (A), sporangia production (B), and virulence (C) of NPTII-harboring vector transformants (Vector) or transformants of P. capsici harboring vector and pcamiR1 antagomir sRNA or control sRNA (CK-sRNA). Colony diameter was measured (A) and photos were taken (upper panel in E) after P. capsici was grown on V8 medium for 3 days. Sporangia were numerated from five randomly selected fields of view under a microscope for three different isolates of each type of transformant (B), and the enumeration and imaging (middle panel in E) were performed 5 days after Vector or CK-sRNA isolates were grown on V8 and 10 days after pcamiR1 antagomir isolates were grown on V8. Scale bar: 50 μm. Mycelial plugs were used to inoculate detached leaves of Nicotiana benthamiana, and lesion size was measured (C) and photos were taken (lower panel in E) at 3 dpi under UV to better visualize the lesions. All the data in A, B, and C are presented as the mean ± SD. Statistical significance compared to the vector and vector + CK-sRNA was determined using Student’s t-test (* P < 0.05, *** P < 0.001, **** P < 0.0001). (D and F) Sporangia development was severely impaired after in vitro application of pcamiR1 antagomir sRNA on P. capsici. Sporangia were numerated from 10 randomly selected fields of view under a microscope for wild-type P. capsici LT1534 after treatment with water, 200 nM control sRNA (CK-sRNA) or pcamiR antagomir sRNA. Enumeration (D) and imaging (F) were performed 6 days after treatment in liquid V8. Scale bar: 200 μm. Data are presented as mean ± SD. Statistical significance compared to water and CK-sRNA treatment was determined using Student’s t-test (* P < 0.05).
Fig 9.
The microRNA-CDK7-RNPII module is conserved.
(A) pcamiR1 is conserved in Phytophthora. PcamiR1 was detected in hypha (HY) or sporulated hypha (HYSP) stage of Phytophthora capsici, Phytophthora infestans, and Phytophthora sojae, respectively. The expression level of P. capsici in the HY stage was given a value of 1.0 and relative gene expression level was calculated by the 2-ΔΔCt method. The housekeeping gene 5S rRNA were used as internal controls to quantify pcamiR1. (B) Sporangia development was severely impaired after in vitro application of pcamiR1 antagomir sRNA on P. infestans and P. sojae. Sporangia were numerated from 10 randomly selected fields of view under a microscope after treatment with 200 nM control sRNA (CK-sRNA) or pcamiR antagomir sRNA. Data are presented as mean ± SD. Statistical significance compared to CK-sRNA treatment was determined using Student’s t-test (* P < 0.05). (C) pcamiR1 analogues were identified in different species belong to animalia. The pcamiR1 analogue miR-365a-5p and miR-365-1-5p were identified in 14 species by searching in miRBase 22.1. A part of the miRNAs are shown and the identical sequences among different miRNAs are highlighted. hsa: Homo sapiens; oan: Ornithorhynchus anatinus; bta: Bos taurus; gga: Gallus gallus. (D) Base pairing and predictive binding energy of pcamiR1 or has-mir-365a-5p with their predicted target sites in hsCDK7 in H. sapiens, respectively. (E and F) CDK7 expression and RNPII Ser5 phosphorylation were inhibited by pcamiR1 and hsa-mir-365a-5p. The relative intensity of hsCDK7, Ser5-phosphorylated RNPII, and unphosphorylated RNPII was quantified 48 h after transfection by ImageJ (E). The control miRNA treated samples detected by each antibody was given a value of 1.00. Data presented in F are as the mean ± SD of three replicates. Statistical significance compared to the control miRNA sample was determined using Student’s t-test (* P < 0.01). β-tubulin was used as the loading control and its intensity in each sample was used to normalize the data between samples.
Fig 10.
The model of RNPII phosphorylation modulation to ensure proper transcription and development.
Sporangia are indispensable for the development and pathogenesis of P. capsici. P. capsici produces sporangia in medium or infected plants after the vegetative growth of hyphae. During the transition from hypha growth to sporangia formation, the transcriptional profile changes dramatically, largely depending on the phosphorylation states of RNPII. During this morphological transition, the cyclin-dependent kinase PcCDK7 is highly induced and promotes the phosphorylation of Ser5 in the characteristic YSPTSPA repeats located in RNPII CTD. As a feedback regulatory machinery, the novel DCL1-dependent microRNA pcamiR1 is also upregulated in line with PcCDK7 and further silences PcCDK7 through complexing with PcAGO1, which ensures the moderate level of PcCDK7. The pcamiR1-PcCDK7 module precisely modulates the phosphorylation state of RNPII and ensures appropriate gene transcription. Therefore, sporangia formation can only be accomplished by accurate gene expression, and any disturbance of the balance between pcamiR1 and PcCDK7 could result in an improper phosphorylation state of RNPII and subsequently lead to the impairment of the normal growth of P. capsici. Due to the conservation of this module in Phytophthora and animalia, pcamiR1 may be a promising target for disease control in the future. Figure created using Figdraw (www.figdraw.com).