Dynamic regulation of genome-wide pre-mRNA splicing and stress tolerance by the Sm-like protein LSm5 in Arabidopsis

Pre-messenger RNA splicing is one of the most consequential steps in gene expression, and it runs on a surprisingly tight tolerance. After a gene is transcribed, the raw precursor messenger RNA still contains introns — non-coding sequences that have to be cut out precisely so the protein-coding exons can be joined. Get the cut sites wrong by even a few nucleotides, and you produce an aberrant transcript that wastes cellular resources, generates a non-functional or harmful protein, or gets degraded entirely. In plants, faulty splicing is directly linked to stress sensitivity. So the question that motivates this research is a practical one: what controls splice-site precision, especially when a plant is suddenly swamped with stress-induced transcripts that all need processing at once? The molecular machines that do the cutting and joining are called spliceosomes — massive ribonucleoprotein complexes built partly around small nuclear RNAs. Among their conserved core components are the Sm-like proteins, known as LSm proteins, which typically assemble into seven-subunit rings. The nuclear version of this ring, the LSm2 to LSm8 heptamer, directly binds and stabilizes the poly-U tail of U6 small nuclear RNA, forming the core of the U6 small nuclear ribonucleoprotein that sits at the heart of the spliceosome. These proteins were long thought of as structural scaffolds — necessary but passive. The work by Cui, Zhang, Ding, Ali, and Xiong in Arabidopsis argues that this picture is wrong. Their study centers on a gene called SAD1, short for Supersensitive to Abscisic Acid and Drought 1, which encodes the LSm5 subunit. SAD1 was originally found in a genetic screen for regulators of stress responses, and plants lacking it are hypersensitive to salt, drought, and the stress hormone abscisic acid. But the molecular reason was never clear. To find out, the team compared three genotypes using RNA sequencing: wild-type Arabidopsis in the C24 ecotype, sad1 loss-of-function mutants, and SAD1 overexpression plants in which wild-type SAD1 was overexpressed in the sad1 mutant background. They ran these comparisons under normal growth conditions and under 300 millimolar sodium chloride salt stress. The RNA sequencing dataset totaled 164 million reads, with roughly 90 percent aligning uniquely to the TAIR10 reference genome. The analysis was designed to detect alternative splicing events, intron retention, and splice-junction usage genome-wide. What they found when SAD1 is missing is striking. Splice-site selection collapsed across thousands of genes. In the sad1 mutant, the team identified 478 alternative five-prime splice sites and 138 exon-skipping events from five hundred fifty genes that were significantly over-represented compared with wild type, which showed only 133 alternative five-prime sites and 41 exon-skipping events from one hundred seventy-one genes. Under salt stress, the mutant also accumulated 319 aberrant alternative three-prime splice sites versus 142 in wild type. Crucially, these alternative sites were not random — they carried the canonical guanine-uridine and adenine-guanine dinucleotides that define real splice sites, and they were enriched within about ten base pairs of the dominant, correct splice sites. The spliceosome wasn't wandering far; it was slipping to near neighbors. About 20 percent of skipped exons in sad1 coincided with these nearby alternative sites, far above the 0.02 percent expected by chance. The biggest damage showed up as intron retention — introns that simply failed to be removed. Cui and colleagues found four thousand six hundred ten introns from two thousand seven hundred thirty-seven genes significantly retained in sad1 plants. In wild type, the comparable number was 23 introns from 20 genes. On average, around 15 percent of total transcripts from affected genes carried a retained intron in the mutant, and those retained introns were predicted to introduce premature stop codons — meaning the transcripts would either be degraded or translated into truncated, non-functional proteins. The genes hit hardest were not random either. The three thousand three hundred fifty-four genes showing abnormal splicing in sad1 were strikingly enriched for stress and abiotic-response pathways. Key salt-tolerance regulators — including SnRK2.1 and SnRK2.2, SOS2, DREB2A, NHX1, WRKY33, WRKY25, STT3A, CAX1, and RCI2A — all showed increased intron retention in the mutant. These are not peripheral players; they include core components of the salt-stress signaling network. Losing SAD1 doesn't just create generic splicing noise. It preferentially scrambles the transcripts the plant needs most when conditions turn hostile. Now flip the experiment. Overexpressing SAD1 doesn't just rescue the mutant — it actively sharpens splicing beyond wild-type levels. In SAD1 overexpression plants treated with salt, the paper identifies 454 alternative splice-site and exon-skipping events from 434 genes that were significantly absent compared to what appeared in wild type. And 76 introns that were over-represented in salt-treated wild type were significantly absent in salt-treated SAD1 overexpression. In total, five hundred six genes showed decreased alternative splicing or intron retention in SAD1 overexpression under salt stress, and the functional, intron-free transcripts from those genes tended to be up-regulated. Splice-junction read profiles in SAD1 overexpression matched wild-type patterns, and sequence analysis confirmed the suppressed alternative sites were the same proximal, cryptic sites that became activated in the mutant. Turning SAD1 up suppressed the nearby alternatives and locked the spliceosome onto the dominant, correct sites. The SAD1 transcript level in overexpression plants was more than ten times higher than in wild type. The phenotypic results tracked the molecular ones. One-week-old seedlings transferred to media containing one hundred millimolar sodium chloride showed that SAD1 overexpression plants had longer roots than either wild type or sad1 mutants. At two hundred millimolar, root elongation was inhibited across all genotypes. In a leaf-damage assay at two hundred millimolar sodium chloride, two-week-old SAD1 overexpression seedlings had a higher percentage of green leaves after five days — statistically significant at a p-value below 0.01. In soil irrigation experiments ramping up to four hundred millimolar sodium chloride over two weeks, sad1 plants were severely damaged above one hundred fifty millimolar, wild-type plants showed damage at higher concentrations, and SAD1 overexpression plants were not obviously affected even at four hundred millimolar and were taller than wild type throughout. These results support what the authors call a dosage-dependent, dynamic model of splicing control. SAD1 is not just structural — its abundance determines how accurately the spliceosome chooses splice sites. The thermodynamic logic goes like this: spliceosomal assembly exists in a dynamic equilibrium between association and dissociation. When SAD1 is depleted, the U6 small nuclear ribonucleoprotein is destabilized, complex assembly is less reliable, and the spliceosome increasingly settles for proximal alternative sites rather than the dominant correct ones. When SAD1 is in excess, assembly is more stable and the dominant sites consistently win out. Under salt stress, when a sudden influx of stress-induced pre-messenger RNAs burdens the splicing machinery, that stability becomes the difference between a functional stress response and a scrambled one. This reframes LSm proteins as regulatory nodes rather than passive scaffolds. The nuclear LSm2 to LSm8 ring isn't just holding U6 small nuclear RNA together; its abundance is a parameter that sets the operating precision of the whole splicing system. Changing that one parameter — SAD1 levels — shifts splicing outcomes at thousands of loci simultaneously, and those shifts have consequences measurable in root length and leaf survival under salt stress. There's a practical takeaway here for crop biology. Overexpressing a splicing fidelity factor improved salt tolerance without visibly disrupting normal growth. That's a different lever than constitutively overexpressing a stress-response gene, which often comes with developmental costs. Enhancing the accuracy of the splicing machinery itself may be a cleaner route to stress tolerance — improving the processing of stress transcripts when they're needed while leaving the baseline transcriptome largely intact. The open questions are real. How are SAD1 levels themselves regulated during stress? Does the plant modulate this splicing dial as part of the stress response? And does the same dosage-sensitivity principle operate for LSm proteins in other eukaryotes — including organisms where splicing fidelity disorders drive disease? Those questions aren't answered here. What is answered is that a single conserved splicing factor, adjusted up or down, can tune the accuracy of splice-site recognition across an entire genome — and that tuning has consequences that reach from molecular fidelity all the way to whether a plant survives a salt flood. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.