Chromatin retained MUSHER lncRNA integrates ABA and DOG1 signalling pathways to enhance Arabidopsis seeds dormancy

Abstract

Many plant lncRNAs regulate gene expression by binding to chromatin, but how they are retained at the target loci is unclear. We identify a chromatin-localized lncRNA - MUSHER, which activates two parallel regulatory pathways to increase Arabidopsis seed dormancy. MUSHER is upregulated in response to high temperatures, contributing to the induction of secondary dormancy. It promotes DOG1 expression by recruitment of the CPSF complex to enhance the proximal cleavage and polyadenylation at the DOG1 transcript. It also increases ABA sensitivity in seeds by activating PIR1 gene transcription. These genes, located on different chromosomes, are both bound by MUSHER, despite lacking sequence homology. The chromatin association of MUSHER enables the integration of the DOG1- and ABA pathways to adjust seed germination timing. Additionally, MUSHER and other lncRNAs interact with U1 snRNP, which is required for their chromatin localisation, revealing a role for U1 snRNP in plants.

Fig. 1. MUSHER lncRNA acts as a positive regulator of seed dormancy and DOG1 expression.

a Schematic representation of the DOG1-MUSHER genomic region of chromosome 5 in A. thaliana. The direction of transcription is indicated with arrowheads. Grey rectangles represent coding exons and introns with thick black lines. The green arrow represents the newly annotated MUSHER transcriptional unit. Numbers along the x-axis on the top indicate genome coordinates (TAIR10). Triangles represent T-DNA insertions of two T-DNA lines (msh-1 and msh-2). The green vertical lines show the positions of multiple guides for msh_dCas9-1 and msh_dCas9-2 dCas9 lines. The green horizontal line corresponds to the region deleted using Cas9 (∆msh). b Primer walking RT-qPCR analysis across the DOG1 locus. The position of amplicons corresponds to the DOG1 locus representation above. Expression levels in the hen2 mutant were normalised to WT and UBC21. Data are presented as mean values of four biological replicates +/− SD. c Positions of MUSHER transcript ends found in WT seeds using 3’ and 5’ RACE-seq, coloured in dark and light green, respectively. d Expression profiles of DOG1 and MUSHER transcripts during seed maturation in the WT. Expression levels were normalised to UBC21 mRNA and 16 DAP. The data are shown as the average of four biological replicates for DOG1 and MUSHER 11DAP, five for MUSHER 16DAP, and six for MUSHER 8,14DAP ± SD. e DOG1 relative expression levels, musher mutants relative to WT in freshly harvested seeds, normalised to UBC21. *p < 0.05, two-tailed Student’s t-test. Data are the mean of five biological replicates, with error bars indicating the standard deviation. f Freshly harvested seeds of WT and different musher mutants were scored for germination. Germination was defined as radical protrusion and counted 4 d after sowing. Data are the mean of six biological replicates +/− SD (n = 6, *p < 0.05, two-tailed t-test). g DOG1 expression profile during silique development in WT and msh_dCas9-1 plants. Expression levels were normalised to UBC21 and shown at the indicated number of days after pollination (DAP); Data are presented as mean values of three biological replicates for WT and four for msh_dCas9-1 + /− SD. h Subcellular localisation of selected transcripts in maturing WT Arabidopsis seeds. Each transcript was tested in 5 biological replicates represented as separate bars. Black horizontal lines represent the mean of chromatin and nucleoplasmic fractions contribution of the total amount of indicated transcript.

Fig. 2. MUSHER enhances DOG1 proximal polyA site selection.

a Snapshot of 3′RNA-seq reads for DOG1 locus in WT and msh-1 mutant (n = 4). Schematic diagram showing DOG1 gene: exons (grey rectangles), shDOG1 3’UTR region (light grey arrow), lgDOG1 3’UTR region (grey arrow), and MUSHER (green). The regions used to estimate the expression of mRNA isoforms are shown below chromosome coordinates. b Ratio of reads corresponding to short and long DOG1 mRNA isoforms calculated for regions marked on the schematic. Data are the mean of four biological replicates; error bars show standard deviations. c The ratio of shDOG1 and lgDOG1 isoforms in freshly harvested seeds of WT and musher mutants, normalised to UBC21 and relative to WT. Data are presented as mean values of four biological replicates +/− SD (*p < 0.05, two-sided t-test). d CPSF73 occupancy was measured by ChIP-qPCR in WT, msh-1 and msh_dCas9-1 mutants. Data are presented as mean values of three biological replicates +/− SD (*p < 0.05, two-sided t-test). e The ratio of shDOG1 and lgDOG1 isoforms in seeds of WT and msh-1, esp1-2 and double msh-1 esp1-2 mutants, normalised to UBC21 and relative to WT (mean of five biological replicates +/− SD, *p < 0.05, two-sided t-test). The ratio of shDOG1 and lgDOG1 isoforms in seeds of WT and msh-1, esp1-2 and double msh-1 esp1-2 mutants, normalised to UBC21 and relative to WT (mean of 5 biological replicates, *p < 0.05, two-sided t-test). f MUSHER relative expression level in seeds of WT and esp1-2, normalised to UBC21 and relative to WT (mean of five biological replicates +/-SD, ns – not significant, two-sided t-test). g Pol II occupancy was measured by ChIP-qPCR in WT and msh-1 mutant. Data are presented as mean values of three biological replicates +/− SD (*p < 0.05, two-sided t-test). h H3, i H3K4me3 and j H3K27me3 levels were measured by ChIP-qPCR in WT and msh-1 mutant. H3K4me3 and H3K27me3 values were normalised to H3 and to ACTIN7 (AT5G09810) gene body. For all ChIP-qPCR experiments presented in this figure: p-value from the two-tailed t-test, *p < 0.05, ns - not significant. Data are the mean of four biological replicates, with error bars indicating the standard deviation. H3, H3K4me3 and H3K27me3 (h–j) ChIP were done side by side and share NoAB control shown on panel h.

Fig. 3. MUSHER activates DOG1 expression during secondary dormancy induction in response to high temperature.

a Seed germination of WT and different musher mutants after 7 days of secondary dormancy (SD) induction at 35 °C. Germination frequency was scored after 5 days. Data are presented as mean values of four biological replicates +/− SD. *p < 0.05 (t-test two-sided). b Seed germination of WT and cpl1-7 mutant seeds after 7 days of SD induction. Germination was scored after 5 days. Data are presented as mean values of five biological replicates +/− SD. *p < 0.05 (t-test two-sided). c The expression profile of DOG1 was analysed using RT-qPCR in seeds of WT, msh-1 and msh_dCas9-1 subjected to secondary dormancy induction. Data are presented as mean values of four biological replicates +/− SD. UBC21 mRNA was used as a reference gene. d The DOG1 relative expression level in seeds of WT and cpl1-7 mutant, subjected to SD induction. Data normalised to UBC21 and related to WT. Data are presented as mean values of four biological replicates +/− SD. p-value from the two-tailed t-test, *p < 0.05. e DOG1 and f MUSHER expression profile in after-ripened seeds subjected to secondary dormancy induction at 25 °C and 35 °C. Expression was normalised to UBC21 mRNA and dry seed levels. Data are presented as mean values of four biological replicates +/− SD (*p < 0.05, two-sided t-test). g Schematic representation of constructs introduced to transgenic plants. RT-qPCR analysis of luciferase mRNA levels in seeds subjected to secondary dormancy induction under temperatures indicated on the right. LUC expression level measured after 5 days of SD induction, normalised to UBC21 and related to 25 °C for each construct, Data are presented as mean values of four biological replicates +/− SD (*p < 0.05, two-sided t-test). h Long and short DOG1 expression analysis on the 5th day of secondary dormancy induction in WT at temperatures indicated. Data are presented as mean values of four biological replicates +/- SD (p-value from the two-tailed t-test, *p < 0.05). UBC21 mRNA was used as a reference gene.

Fig. 4. MUSHER enhances PIR1 expression independently of its role in DOG1 regulation.

a Volcano plot showing differentially expressed genes for WT and msh-1 selected using DESeq2; absolute log2FC >  log2 (1.5), FDR  <  0.05, n = 4. DOG1 and PIR1 genes are highlighted. The number of downregulated (down) and upregulated (up) genes is provided on the plot. b Diagram of 3′RNA-seq reads for PIR1 locus in WT and msh-1 mutant. Schematic diagram showing PIR1 locus: exons (grey rectangles), 3’UTR region (grey arrow) c The PIR1 relative expression level in freshly harvested seeds (primary dormancy) of WT, msh-1 and dog1-4 mutants, normalised to UBC21 and relative to WT. Data are presented as mean values of four biological replicates +/− SD (p-value from the two-tailed t-test, *p < 0.05). *p < 0.05. d The PIR1 relative expression level in WT msh-1, dog1-4 and cpl1-7 mutant seeds, after 5 days of SD induction, normalised to UBC21 and relative to WT. Data are presented as mean values of four biological replicates +/− SD (p-value from the two-tailed t-test, *p < 0.05). e Pol II f H3 g H3K4me3 and h H3K27me3 occupancy was measured by ChIP-qPCR in WT, msh-1. (n = 4) H3K4me3 and H3K27me3 values were normalised to H3 and to ACTIN7 (AT5G09810); Pol II and H3 were normalised to ACTIN7 (AT5G09810). For all ChIP experiments *P < 0.05, ns - not significant (two-tailed t-test). Error bars represent the standard deviation of four biological replicates. H3, H3K4me3 and H3K27me3 (f–h) ChIP were done side by side and share NoAB control shown on panel h.

Fig. 5. MUSHER lncRNA integrates ABA and DOG1 pathways by interacting with PIR1 TSS and DOG1 proximal termination site.

Snapshot of ChIRP-seq in WT and msh-1 mutant using MUSHER-specific and LacZ-specific probes on a DOG1 and b PIR1 loci. Schematic diagram showing DOG1 and PIR1 locus: exons (grey rectangles), shDOG1 3’UTR region (light grey arrow), lgDOG1 3’UTR region (grey arrow), and MUSHER (green). The lines below the scheme mark regions used to design MUSHER probes. Data are the mean of two biological replicates. c Snapshot of ChIDP-seq in WT using PIR1-specific probes on DOG1 and d PIR1 loci. Data are the mean of two biological replicates. The lines below the scheme mark regions used to design PIR1 probes. e The germination rate of WT, dog1-3, at2g35320-1, pir1 and msh-1 mutant seeds on MS media supplemented with various ABA concentrations. Data are presented as mean values of four biological replicates for seeds germinated on ABA + /− SD and 2 biological replicates for mock. Asterix denotes a two-tailed Welch’s t-test comparing the wild type (WT) to each mutant yielded a p < 0.05.

Fig. 6. U1 is required for chromatin tethering of MUSHER and other lncRNA.

a MUSHER or LacZ probes were used in formaldehyde-assisted RIP to enrich associated RNA. U1, U4 and U6 recovery was analysed using RT-qPCR. Data are the mean of four biological replicates +/− SD. b MUSHER probes were used to enrich MUSHER from WT and as a negative control in msh-1. U1 and U6 recovery was analysed using RT-qPCR. The data represent the average of four biological replicates +/− SD. c U1A protein was precipitated side by side with a no-antibodies (NoAB) control from plants expressing pU1-70K::U1-70K-GFP u1-70k, and co-precipitated RNA was analysed using RT-qPCR for spliced (AT5G45830) and unspliced (AT5G45820) RNA controls and MUSHER. The data are averaged from three biological replicates + /− SD. A two-tailed Welch’s t-test p-value was calculated compared to WT (*p < 0.05). d Subcellular localisation of mRNAs and lncRNAs in maturing Arabidopsis seeds of WT and u1c amiRNA line. Lines represent the mean of chromatin and nucleoplasmic fractions contribution of the total amount of indicated transcript. The blue-shaded colour refers to Col-0, while the grey-shaded colour corresponds to u1c. Each transcript was measured in 4 replicates represented as a separate bar. *p  <  0.01 from two-tailed Student t-test. e After ripened seeds of WT (WT), u1c amiRNA and u1k-70k mutants were induced into secondary dormancy, moved to permissive conditions and scored for germination ability. Germination (%) was scored as radical protrusion. Data are the mean of four biological replicates +/− SD (two-tailed t-test, *p < 0.05). f shDOG1 expression was analysed in WT, u1c amiRNA and u1-70k mutants seeds induced into secondary dormancy. Data are the mean of four biological replicates +/− SD. A two-tailed t-test p was calculated compared to WT (*p < 0.05). Normalised to UBC21 and relative to WT.

Fig. 7. MUSHER regulates seed dormancy and lncRNA chromatin retention in Arabidopsis.

a MUSHER modulates seed dormancy through the transcriptional regulation of DOG1 and ABA pathways. Created in BioRender Halale manjunath, V. (2025) https://BioRender.com/djj6tmh. b U1 is required for chromatin retention of lncRNA, including MUSHER in Arabidopsis. Created in BioRender. Halale manjunath, V. (2025) https://BioRender.com/zss2438.