Plants respond to environmental stresses through controlled stem cell maintenance and meristem activity. One level of transcriptional control is RNA alternative splicing. However the mechanistic link between stress, meristem function and RNA splicing is poorly understood. The MERISTEM-DEFECTIVE (MDF)/DEFECTIVELY ORGANIZED TRIBUTARIES (DOT2) gene of Arabidopsis encodes a SR-related family protein, required for meristem function and leaf vascularization, and is the likely orthologue of the human SART1 and yeast snu66 splicing factors. MDF is required for the correct splicing and expression of key transcripts associated with root meristem function. We identified RSZ33 and ACC1, both known to regulate cell patterning, as splicing targets required for MDF function in the meristem. MDF expression is modulated by osmotic and cold stress, associated with differential splicing and specific isoform accumulation and shuttling between nucleus and cytosol, and acts in part via a splicing target SR34. We propose a model in which MDF controls splicing in the root meristem to promote stemness and repress stress response and cell differentiation pathways.

RNA extraction and sequencing

RNA was extracted from three independent biological replicates using 7-day-old seedlings (ca. 100 mg tissue) grown on half-strength MS10 medium using the Sigma-Aldrich Plant Total RNA Kit (catalog number STRN50), with the On-Column DNase I Digestion Set (catalog Number: DNASE10-1SET) to eliminate any residual DNA molecules. Plant tissue was ground in liquid nitrogen before incubation in a lysis solution containing 2-mercaptoethanol at 65°C for 3 minutes. The solid debris was removed by centrifuging and 14 000 x g and column filtration before RNA was captured onto a binding column using the supplied binding solution, which helps preventing polysaccharide and genomic DNA from clogging the column. Most DNA was removed by wash solutions, and any trace of residual DNA was removed by DNase on the column. Then purified RNA was eluted using RNAase-free water.

RNA sequencing from three biological replicate samples was carried out on an Illumina HiSeq 2500 System with the library prepared using the Illumina TruSeq Stranded Total RNA with Ribo-Zero Plant Sample Preparation kit (catalog Number: RS-122-2401). Ribosomal RNA (rRNA) was removed from isolated total RNA using biotinylated, target-specific oligos on rRNA removal beads. Purified RNA was quality checked using a TapeStation 2200 (Agilent Technology) with High Sensitivity RNA ScreenTape (catalog Number: 5067-5579), and the mRNA was fragmented into 120-200 bp sequences with a median size of 150 bp. Fragmented mRNA was used as a template to synthesise first-strand cDNA using reverse transcriptase and random primers, followed by second-strand cDNA synthesis with DNA Polymerase I and RNase H. Newly synthesised cDNA had a single adenine base added with ligation of adaptors, before being purified and amplified by PCR to make the final library. Library quality control was performed again using a TapeStation with D1000 ScreenTape (catalog Number: 5067-5582).

Pre-processing of RNA-seq data, differential expression and differential usage analysis.

RNAseq data were processed and aligned against the TAIR10 (EnsemblePlants) genome using TopHat and indexed with Samtools. DeSeq determined differential expression. Alternative splicing analysis was determined using RMats (p value of 0.05, a minimum of 10% inclusion difference). Alternative splicing events were visualised using Sashimi plots generated by the Integrative Genomics Viewer (IGV) (Robinson et al., 2011).

Direct mRNA isolation and cDNA preparation for RT-qPCR or RT-PCR

Seedlings were grown 7 days post-germination as described above. Roots and cotyledons were separated using a razor blade, and the material was frozen immediately in liquid nitrogen. Pools of seedlings were used to generate three separate biological samples. Each pool contained approximately 20 mg of root or cotyledon tissue. Total mRNA was extracted using Dynabeads®mRNA DIRECT™kit with Oligo(dT)25 labelled magnetic beads. Frozen tissue was ground with a sterile plastic micropestle and resuspended in 300 µl lysis buffer. The solution was then forced through a 21-gauge needle in a 1ml syringe 3-5 times to shear any DNA and mixed with 50 µl of Dynabeads Oligo(dT)25. The kit procedure was followed, with two final washes conducted. To ensure the complete removal of any genomic DNA in the subcellular fractionation experiments, this stage was followed by ezDnase™ treatment in a 10 µl volume (1µl ezDNASe™, 1 µl ezDNASe™ 10X buffer and 8 µl sterile H2O), 37˚C for 2 minutes followed by 1 µl DTT and 5 minutes at 55˚C in a heat block.

cDNA was prepared using a SuperScript®IV First-Strand synthesis system directly on the bead solution. For RT-PCR and RT-qPCR beads were washed in 20 µl 1 x SSIV buffer before resuspension in 12 µl sterile H2O with 1 µl dNTP 10 mM each mix and incubated for 5 minutes at 50 ˚C in a Proflex PCR machine (Applied Biosystems). Then the following were added 4 µl 10 x SSIV buffer, 1 µl ribonuclease inhibitor and 1 µl Superscript®IV reverse transcriptase were added. The mixture was mixed by pipetting and incubated for 10 minutes at 50 ˚C, followed by 10 minutes at 80 ˚C, and then held at 4 ˚C. The 20 µl cDNA mix was stored at -20 and not eluted from the beads.

Samples were checked for the presence of genomic DNA by PCR with Actin 2 primers ACT2 forward and reverse. A PCR reaction after 28 cycles with a Tm of 60 ˚C generated a 340 bp product if genomic DNA was a contaminant, 240 bp otherwise. All PCR and sequencing primers are listed in Table S3.

RT-PCR

0.5 – 1 µl of cDNA/bead mix were used per PCR reaction. RT-PCR was performed with RSZ33 or ACC1 root-derived cDNA using Phusion™ (Thermofisher) high-fidelity polymerase. Relative levels of RSZ33 and ACC1 splice variants were determined using FIJI gel analysis software (Schindelin et al. 2012). Relative levels of cDNA per sample were determined using PCR-amplified PP2a transcript levels. 

Excel files are provided.