Volume 8, Issue 7, Pages 1053-1068 (July 2015) The Splicing Factor PRP31 Is Involved in Transcriptional Gene Silencing and Stress Response in Arabidopsis  Jin-Lu Du, Su-Wei Zhang, Huan-Wei Huang, Tao Cai, Lin Li, She Chen, Xin-Jian He  Molecular Plant  Volume 8, Issue 7, Pages 1053-1068 (July 2015) DOI: 10.1016/j.molp.2015.02.003 Copyright © 2015 The Author Terms and Conditions

Figure 1 The prp31 Mutation Reactivates the Expression of the RD29A-LUC Transgene in the ros1/ago4 Mutant Background. (A) The RD29A-LUC transgene expression was reflected by luminescence imaging in the C24 wild-type, ros1, ros1/ago4, and ros1/ago4/#156 after cold treatment (4°C for 48 h). (B) The diagram of map-based cloning of the prp31 mutation on chromosome 1 (Chr. 1). The mutation was mapped to an 850-kb interval on the bottom arm of chromosome 1. A nucleotide substitution (G to A) was identified in the last introns of PRP31 (At1G60170). PRP31 contains a Nop domain and a Prp31 C-terminal domain. (C and D) The PRP31-Myc transgene complements the luminescence and developmental phenotypes of ros1/ago4/prp31. Molecular Plant 2015 8, 1053-1068DOI: (10.1016/j.molp.2015.02.003) Copyright © 2015 The Author Terms and Conditions

Figure 2 The prp31 Mutation Affects the Transcriptional Silencing of Transgenic and Endogenous RD29A in a DNA Methylation-Independent Manner. (A) Luminescence imaging of indicated genotypes after cold treatment (4°C for 48 h). (B) Morphology phenotypes of 3-week-old C24 and indicated mutants. (C) The RNA transcript levels of LUC, RD29A, and NPTII were measured by quantitative RT–PCR. Total RNA was extracted from 10-day-old seedlings of the C24 wild-type (WT), prp31, ros1, ros1/prp31, ros1/ago4, and ros1/ago4/prp31 with or without cold treatment (4°C for 48 h). The expression of ACT2 was used as an internal control. Error bars represent the standard deviations. Quantitative RT–PCR experiments were independently repeated three times with similar results, and data from one repetition are shown. (D) DNA methylation of transgenic and endogenous RD29A promoter regions was determined by bisulfite sequencing. The percentages of cytosine methylation in CG, CHG, and CHH sites are indicated. Molecular Plant 2015 8, 1053-1068DOI: (10.1016/j.molp.2015.02.003) Copyright © 2015 The Author Terms and Conditions

Figure 3 PRP31 Associates with ZOP1 and STA1 in the Nucleus. (A–C) PRP31 directly interacts with ZOP1 and STA1 but not AGO4 in vivo. PRP31-Myc was coimmunoprecipitated with STA1-Flag and ZOP1-Flag in the corresponding F1 transgenic plants expressing both epitope-tagged transgenes. AGO4 was not detected from the immnunoprecipitation of PRP31-Flag. The Flag- and Myc-tagged proteins and AGO4 were detected by western blotting. WT, wild-type. (D) PRP31 was specifically localized in the nucleus as determined by western blotting. UGPase and histone H3 were used as markers of the cytosolic (C) and nuclear (N) fractions, respectively. T represents total protein extracts. (E) Immunolocalization of PRP31-Myc and ZOP1-Flag in the nucleus of the F1 plants expressing both PRP31-Myc and ZOP1-Flag transgenes. The DAPI-stained blue signals visualized the nucleus; the green signals and red signals visualized the distribution of ZOP1-Flag and PRP31-Myc, respectively, in the nucleus. Yellow signals were generated if the two epitope-tagged proteins colocalized. The percentages of nuclei with indicated signal patterns are shown to the right. (F) Immunolocalization of STA1 and PRP31-Myc in the nucleus of the PRP31-Myc transgenic plants. (G) Immunolocalization of U2B and PRP31-Myc in the nucleus of the PRP31-Myc transgenic plants. Molecular Plant 2015 8, 1053-1068DOI: (10.1016/j.molp.2015.02.003) Copyright © 2015 The Author Terms and Conditions

Figure 4 The Function of PRP31 in Transcriptional Gene Silencing Is Independent of the RNA-Directed DNA Methylation Pathway. (A) RNA transcript levels of the RdDM targets AtGP1, solo LTR, SDC, and TUB4 were analyzed by quantitative RT–PCR in the C24 wild-type (WT), prp31, ago4, and ago4/prp31. TUB4 was taken as an unchanged control locus between C24 and mutants. (B) DNA methylation of AtSN1, IGN5, IGN23, and solo LTR loci was assessed by chop–PCR. (C) Northern blot analysis of small RNAs in the C24 wild-type, ros1, ros1/prp31, ros1/ago4, ros1/ago4/prp31, and ros1/nrpd1. The ethidium bromide-stained RNA was used as a loading control. (D) The genetic relationship between prp31 and the RdDM mutations was determined by the expression of RD29A-LUC. Luminescence signals indicate the expression of RD29A-LUC after cold induction (4°C for 48 h). Molecular Plant 2015 8, 1053-1068DOI: (10.1016/j.molp.2015.02.003) Copyright © 2015 The Author Terms and Conditions

Figure 5 Effect of prp31, zop1, and sta1 on the Expression of Protein Coding Genes and Transposable Elements as Determined by RNA Deep Sequencing. (A) Heat-map showing the overall up- and down-regulated protein coding genes in these splicing factor mutants (fold-change ≥2, P ≤ 0.01; fold-change ≤0.5, P ≤ 0.01). The value represents the fold-change of mutant reads versus the C24 wild-type on a logarithmic scale (log2). (B) Venn diagrams show the relationship between groups of up- or down-regulated protein coding genes in different splicing factor mutants (fold-change ≥2, P ≤ 0.01; fold-change ≤0.5, P ≤ 0.01). The Roman numbers on the right indicate the subsets tested in the validation experiments. (C) The RNA expression levels of the up- or down-regulated genes were validated by semiquantitative RT–PCR. (D) Venn diagrams show the relationship between groups of up- or down-regulated transposable element genes in different splicing factor mutants (fold-change ≥2, P ≤ 0.05; fold-change ≤0.5, P ≤ 0.05). (E) The up-regulation of the indicated transposable elements (TEs) in prp31 was confirmed by quantitative RT–PCR. ACT2 was used as an internal control. The results in (C) and (E) were repeated for three times with similar results, and data from one experiment are shown. Molecular Plant 2015 8, 1053-1068DOI: (10.1016/j.molp.2015.02.003) Copyright © 2015 The Author Terms and Conditions

Figure 6 PRP31 Is Required for Pre-mRNA Splicing. (A) The splicing defects in prp31 at three loci were highlighted with frames according to the RNA deep sequencing data. Black bars, gray bars, and lines represent exons, untranslated regions, and introns, respectively. The normalized reads coverage in the C24 wild-type and the prp31 mutant are shown in wiggle plots. (B) Validation of the intron retention events by quantitative RT–PCR. Primers flanking the 3′ splicing site were used to detect introns as illustrated, and error bars represent standard deviations (n = 3). Quantitative RT–PCR was repeated three times with similar results, and data from one experimental repetition are shown. Molecular Plant 2015 8, 1053-1068DOI: (10.1016/j.molp.2015.02.003) Copyright © 2015 The Author Terms and Conditions

Figure 7 PRP31, ZOP1, and STA1 Are Required for Proper Seed Germination under Various Stress Conditions. (A) Germination of the wild-type C24 and the prp31, zop1, and sta1 mutants on half-strength MS agar medium with supplements. The plates were photographed after 8 days in the growth chamber. A seed was scored as germinated if it produced green cotyledons, and representative seeds are marked with a red circle. (B) Germination percentages of the wild-type and prp31 on half-strength MS agar medium with supplements. The germination percentages were scored after 7 days of incubation. Values are means of three replicates, with n ≥ 30 per replicate. Molecular Plant 2015 8, 1053-1068DOI: (10.1016/j.molp.2015.02.003) Copyright © 2015 The Author Terms and Conditions

Figure 8 PRP31 Is Important for Pre-mRNA Splicing, Cold-Responsive Gene Regulation, and Cold Tolerance. (A) Chilling sensitivity test of the wild-type C24, prp31, zop1, and sta1 at 4°C in darkness. Up, 23°C for 7 days; down, 4°C for 21 days. The white line represents 1 cm. (B) Quantitative RT–PCR analysis of the cold-responsive genes in 10-day-old wild-type (WT) and prp31 seedlings with or without cold treatment (4°C for 48 h). (C) Quantitative RT–PCR analysis of the mature and nonspliced transcripts of RD29A, COR6.6, and COR15A. Primers F1 and R1 were designed from two adjacent exons; primers F2 and R2 were designed in the intron as illustrated. The cold treatment procedure was the same as in (B). Similar results were obtained from three independent repeats, and data from one repetition are shown. Molecular Plant 2015 8, 1053-1068DOI: (10.1016/j.molp.2015.02.003) Copyright © 2015 The Author Terms and Conditions