Heat stress damages plants through several interacting routes. High temperatures can destabilize membranes, impair proteins, increase reactive oxygen species, and disrupt metabolic pathways. Plants counter these effects through signaling networks that activate heat-shock proteins, transcription factors, protective metabolites, and repair systems. Much of the research on thermotolerance has focused on transcription: which genes are turned on or off. But a gene can produce multiple mature messenger RNAs, and these molecules are not necessarily equivalent. During pre-mRNA processing, cleavage and polyadenylation factors select a site near the RNA molecule’s 3-prime end and add a polyadenosine tail. When a gene contains several possible sites, the process is called alternative polyadenylation, or APA. A proximal site creates a shorter RNA, whereas a distal site generally preserves a longer 3-prime untranslated region. The resulting transcripts may differ in stability, translation, regulatory interactions, or protein-coding capacity.

The researchers examined whether SIZ1, already known to support basal heat tolerance in Arabidopsis, also controls APA during acute heat stress. SIZ1 attaches small ubiquitin-like modifier proteins, known as SUMOs, to target proteins through a post-translational modification called SUMOylation. Earlier work had implicated SIZ1 in heat-responsive transcription and in APA during thermomorphogenesis, the growth changes plants make under warm conditions. Extreme heat stress, however, is biologically distinct from mild warmth, so its effects on RNA 3-prime end formation required separate investigation. The team compared normal Col-0 Arabidopsis seedlings with the siz1-2 mutant, which lacks functional SIZ1 activity. Seven-day-old seedlings were exposed either to 22 degrees Celsius or to 37 degrees Celsius. For the genome-wide experiment, the heat treatment lasted 30 minutes, allowing the researchers to capture early changes in polyadenylation patterns.

To map RNA endings, the researchers used poly(A) tag sequencing, or PAT-seq, a method designed to identify polyadenylation sites across the transcriptome. The analysis detected 49,063 poly(A) site clusters associated with 18,229 genes. About 67 percent of those genes used more than one polyadenylation site and therefore qualified as APA genes. More than 60 percent of the identified site clusters and over 40 percent of the sequencing tags fell within 3-prime untranslated regions, although heat stress also altered sites in exons, introns, extended untranslated regions, and intergenic regions. Compared with untreated seedlings, heat exposure changed 1,719 polyadenylation sites in Col-0 and 2,202 in the siz1-2 mutant. More than 900 transcripts were specifically regulated by SIZ1, indicating that the SUMO ligase affects a broad collection of RNA products rather than a small set of isolated genes.

The team then separated genes with altered polyadenylation from genes whose overall expression changed. Heat stress altered more than 700 differentially expressed APA genes, including 189 whose APA changes depended on SIZ1. In the mutant, loss of SIZ1 changed the expression of 480 APA genes. Gene ontology analysis linked upregulated APA transcripts in normal seedlings to responses to heat and temperature stimuli, while pathway analysis connected other changes to plant hormone signaling, amino-acid biosynthesis, and chlorophyll metabolism. These results suggest that APA is embedded in several physiological systems affected by high temperature. The researchers also used a weighted clustering approach to identify switch genes, defined as genes that changed the relative use of their alternative polyadenylation sites. More than 300 such genes showed heat-associated expression changes, and SIZ1 specifically regulated 125 switch genes when the mutant and normal plants were compared under heat.

Many of the switches occurred in canonical 3-prime untranslated regions, but the study also detected changes involving non-canonical sites in intragenic regions outside annotated 3-prime untranslated regions. The distinction matters because a longer or shorter untranslated region can modify RNA behavior without changing its protein-coding sequence, whereas cleavage in an upstream or unusual region can produce a truncated or otherwise different protein. Under heat stress, the plants frequently shifted toward distal 3-prime untranslated region sites, a pattern particularly evident in the SIZ1-deficient background. The researchers also found differences in the nucleotide signals surrounding the selected sites. Adenine-rich elements near the cleavage site and uracil-rich motifs farther upstream varied between lengthened and shortened transcripts, suggesting that heat-responsive site choice is influenced by the sequence features recognized by the RNA-processing machinery. The data support a model in which heat stress reshapes both standard and non-canonical RNA endings.

To test whether individual transcript forms had distinct biological effects, the researchers focused on four heat-responsive genes. DREB2A and HSFA3 encode transcription factors central to heat responses, and their distal transcripts promoted expression of downstream heat-shock genes in cell-based experiments. The team also examined GolS2, which encodes galactinol synthase and participates in the production of raffinose-family oligosaccharides, and TTL3, a tetratricopeptide repeat-like protein associated with stress-related molecular complexes. Because proximal transcripts are difficult to measure specifically—the distal RNA contains the sequence found in the proximal form—the researchers designed reverse primers carrying a transcript-end-specific sequence paired with a polyadenosine tract. Reverse transcription quantitative PCR confirmed that the selected primers distinguished the proximal and distal isoforms of TTL3 and DREB2A. Heat exposure increased the proximal-to-distal ratio for HSFA3, GolS2, and TTL3, but decreased it for DREB2A.

The strongest functional evidence came from plants engineered to overexpress individual transcript variants. Arabidopsis seedlings carrying the distal GolS2 transcript survived heat treatment more effectively than wild-type seedlings, whereas overexpressing the proximal form did not produce the same benefit. The opposite pattern appeared for TTL3: the proximal transcript enhanced heat tolerance, while the distal transcript performed similarly to the wild type. The experiment exposed seven-day-old seedlings to 37 degrees Celsius for four days, followed by three days of recovery at 22 degrees Celsius. In the SIZ1 mutant background, the protective GolS2 distal isoform restored the heat-induced expression of HSP18 and HSP22, while the alternative form did not. For TTL3, the proximal isoform rescued the expression of those heat-shock genes. Protein measurements offered a possible explanation: the proximal TTL3 transcript produced more TTL3 protein, while the two GolS2 transcripts generated protein forms with distinct sizes and different abundance patterns under heat.

The results establish a connection between SUMOylation, RNA 3-prime end processing, and plant thermotolerance. Additional assays indicated that heat stress increased SUMOylation of CPSF100, a component of the cleavage and polyadenylation specificity factor complex, and that this modification depended on SIZ1. The finding suggests that SIZ1 may regulate APA by modifying a core RNA-processing factor, thereby helping the cell select transcript endings during heat exposure. The work does not yet explain precisely how each alternative 3-prime end changes RNA stability, translation, or protein activity, and the experiments were conducted in Arabidopsis seedlings rather than crop plants under field conditions. Nevertheless, the study offers a practical strategy: genome-wide APA maps can reveal candidate transcript isoforms, which can then be tested individually for stress-protective functions. Because APA is widespread among eukaryotes, the approach may eventually help researchers investigate heat resilience in other plants and identify molecular targets for crop improvement.

The study’s experimental design separates rapid RNA-processing responses from longer-term survival outcomes. PAT-seq was performed after only 30 minutes at 37 degrees Celsius, whereas the recovery assay involved four days of heat followed by three days at 22 degrees Celsius. This distinction is important: the sequencing experiment captures early changes in polyadenylation-site usage, while the later phenotype reflects the cumulative effects of protein protection, metabolism, cellular repair, and developmental recovery. The results therefore suggest that altered transcript endings arise early enough to contribute to downstream heat adaptation, although the study does not establish a complete causal timeline for every isoform.

The findings also illustrate why measuring total gene expression can miss biologically important regulation. Two transcripts from one locus may be counted together in a conventional RNA-sequencing analysis even when their 3-prime ends confer different regulatory properties. In this work, the researchers used transcript-end-specific reverse-transcription quantitative PCR to distinguish selected proximal and distal products, addressing a technical problem created by the sequence overlap between isoforms. That strategy is especially relevant for APA studies because a distal transcript contains much of the sequence present in a shorter proximal transcript, making nonspecific measurements likely to overestimate the abundance of the shorter form.

At the mechanistic level, the proposed link to CPSF100 places SIZ1 close to the core machinery that recognizes cleavage and polyadenylation signals. The authors report that heat increased CPSF100 SUMOylation in a SIZ1-dependent manner, but the evidence does not yet show whether this modification directly changes CPSF100’s binding preferences, its interaction with other processing factors, or the timing of cleavage. Nor does it determine whether each protective isoform acts mainly through altered RNA lifetime, translation, or protein structure. These questions will require measurements of transcript stability, ribosome association, protein activity, and site-specific processing in additional tissues and developmental stages. Such validation will be necessary before the candidate isoforms can be evaluated for usefulness in crop breeding or engineering, where heat exposure is repeated, variable, and combined with other environmental stresses.

Subject of Research: SIZ1-dependent alternative polyadenylation in Arabidopsis heat stress tolerance

Article Title: Identification of specific transcripts for plant heat stress tolerance through genome-wide analysis of SIZ1-dependent alternative polyadenylation

Article References: Wang, J., Wu, X., Zhou, Z., Zheng, S., Hu, M., Xiao, Y., Shi, L., Zhang, C., Li, J., Yang, C., Lai, J., Han, D., & Yu, Z. (2026). Identification of specific transcripts for plant heat stress tolerance through genome-wide analysis of SIZ1-dependent alternative polyadenylation. Stress Biology, 6(1), Article 61. https://doi.org/10.1007/s44154-026-00340-2

Image Credits: AI Generated

DOI: 10.1007/s44154-026-00340-2

Keywords: Heat stress, Alternative polyadenylation, SIZ1, Arabidopsis, GolS2, TTL3, RNA processing, Plant thermotolerance, Identification, specific, transcripts, plant