At-RS31 orchestrates hierarchical cross-regulation of splicing factors and integrates alternative splicing with TOR-ABA pathways

Abstract

Introduction

Gene expression in eukaryotes involves multiple regulatory layers. Following transcription, nascent RNAs undergo processing steps, including capping, splicing, polyadenylation, and chemical modification, to produce mature mRNAs (Yang et al., 2021). Splicing removes introns from pre-mRNA and joins exons to generate the mature transcripts (Gilbert, 1978). Although splicing is a highly regulated process ensuring specificity, it also shows remarkable plasticity. The spliceosome, the cellular machinery responsible for splicing, can recognize alternative splice sites, enabling a single gene to produce multiple transcript variants via alternative splicing.

In plants, 40–70% of intron-containing genes undergo alternative splicing, underscoring its fundamental role in regulating gene expression during development and environmental responses (Filichkin et al., 2010; Lu et al., 2010; Marquez et al., 2012; Chamala et al., 2015). Alternative splicing not only produces diverse transcripts leading to different proteins but also generates noncoding isoforms, which may be rapidly degraded or remain stable, thus fine-tuning the total protein levels produced by a gene (Kalyna et al., 2012; Petrillo, 2023). Different types of alternative splicing events, such as exon skipping (ES), intron retention, and usage of alternative 5′ and 3′ splice sites, generate transcript diversity. While ES is common in animals, intron retention is most frequent in plants. Retained intron (RI) transcripts often remain in the nucleus, regulating protein levels during stress or developmental transitions (Kalyna et al., 2012; Marquez et al., 2012; Yap et al., 2012; Boothby et al., 2013; Leviatan et al., 2013; Braunschweig et al., 2014; Gohring et al., 2014; Boutz et al., 2015). Furthermore, exitrons (EIs), alternatively spliced internal regions within protein-coding exons, add another layer of complexity to the alternative splicing landscape (Marquez et al., 2015; Staiger & Simpson, 2015).

The spliceosome ensures the accurate recognition of different pre-mRNA regions and intron removal, aided by numerous proteins. Among these proteins, two key groups stand out: serine/arginine-rich (SR) proteins and heterogeneous nuclear ribonucleoproteins (hnRNPs; Wachter et al., 2012). Serine/arginine-rich proteins interact with the pre-mRNA and spliceosomal components, guiding spliceosome assembly at specific splice sites (Shepard & Hertel, 2009). They contain one or two N-terminal RNA recognition motifs (RRMs), the most prevalent RNA-binding domain, and a C-terminal arginine/serine (RS) region enriched in arginine/serine dipeptides, which engages primarily in protein–protein interactions but also contributes to RNA recognition. By interacting with the spliceosomal machinery, SR proteins modulate splice site selection, contributing to mRNA isoform diversity. Beyond splicing, SR proteins influence transcription (Lin et al., 2008; Ji et al., 2013), polyadenylation (Schwich et al., 2021), mRNA export (Müller-McNicoll et al., 2016; Botti et al., 2017), translation (Sanford et al., 2004) among other processes. However, most knowledge about SR protein functions comes from animal studies.

In plants, the SR protein family has expanded remarkably, with Arabidopsis thaliana possessing 18 SR proteins classified into six subfamilies. Ten are plant-specific, divided into RS, RS2Z, and SCL subfamilies based on their domain organization. The remaining eight are similar to mammalian SR proteins SF2/ASF/SRSF1, 9G8/SRSF7, and SC35/SRSF2 and belong to SR, RSZ, and SC subfamilies, respectively (Kalyna & Barta, 2004; Barta et al., 2008, 2010; Duque, 2011; Richardson et al., 2011). Arabidopsis also has two SR-like proteins: SR45 and SR45a. These proteins participate in constitutive and alternative splicing and play roles in mRNA export, stability, translation, transcriptional elongation, and cell cycle regulation (Jin, 2022). Several SR proteins contribute to plant development and abiotic stress responses (Jin, 2022), but their in vivo targets and regulatory networks remain less characterized (Mateos & Staiger, 2023). So far, RNA immunoprecipitation (RIP) followed by RNA sequencing (RNA-seq) identified over 4000 RNAs associated with SR45 in Arabidopsis seedlings (Xing et al., 2015) and 1812 in inflorescences (Zhang et al., 2017). Recently, tomato RS2Z35 and RS2Z36 were shown to bind to transcripts of over 5000 genes, including the heat shock transcription factor (TF) and many transcripts that undergo heat shock-sensitive alternative splicing, preferentially binding purine-rich RNA motifs (Rosenkranz et al., 2024).

At-RS31 (AT3G61860), a plant-specific SR protein in the RS subfamily, may regulate unique plant functions, although its exact roles are unclear. It has two N-terminal RRMs and the RS region, specific to this subfamily (Supporting Information Fig. S1a; Lopato et al., 1996). At-RS31 interacts with SR and SR-like proteins and spliceosome components, suggesting a role in pre-mRNA splicing (Lopato et al., 2002; Lorkovic et al., 2005; Altmann et al., 2020). Its ability to stimulate splicing in SR protein-deficient HeLa cell S100 extracts further supports this role (Lopato et al., 1996).

At-RS31 undergoes alternative splicing, producing four transcript isoforms: mRNA1-4 (Lopato et al., 1996; Fig. S1b,c). The shortest isoform, mRNA1, arises from excision of the entire intron 2 and encodes the SR protein. This whole intron is retained in mRNA4. mRNA3 uses a proximal 3′ splice site in intron 2, while mRNA2 arises either from the removal of a small intron in mRNA3 or from the inclusion of a cassette exon compared with mRNA1. mRNA2–4 contain premature termination codons (PTCs). Only mRNA2 is sensitive to nonsense-mediated mRNA decay (NMD), likely due to nuclear retention of mRNA3 and mRNA4 (Kalyna et al., 2012; Petrillo et al., 2014).

The ratio of At-RS31 isoforms varies with tissue type, developmental stage, and environmental stimuli, such as bacterial flagellin, cold, or red light (Lopato et al., 1996; Palusa et al., 2007; Tognacca et al., 2019; Bazin et al., 2020). The proportion of mRNA1 fluctuates in response to light, increasing in light and decreasing in darkness, a response mediated by chloroplast retrograde signalling, affecting even nonphotosynthetic root cells lacking chloroplasts; sugars, mitochondrial function, and the Target of Rapamycin (TOR) pathway are key to this effect (Petrillo et al., 2014; Riegler et al., 2021). The conserved alternative splicing pattern of At-RS31 across diverse plant species, from green algae to flowering plants, underscores its biological significance (Iida & Go, 2006; Kalyna et al., 2006).

Despite the identification of many factors regulating At-RS31 alternative splicing, hinting at its potential roles in several plant biological processes, its downstream targets remain unknown. Since SR proteins influence alternative splicing in a concentration-dependent manner (Mayeda et al., 1992) and given the dynamic modulation of At-RS31 in response to various environmental and developmental signals, we hypothesize that its expression levels significantly impact the transcriptome. To identify direct targets of At-RS31, we performed in vivo individual-nucleotide resolution crosslinking and immunoprecipitation (iCLIP; Konig et al., 2010; Hafner et al., 2021) followed by in vitro RNAcompete validation (Ray et al., 2013, 2017). Additionally, we performed transcriptome profiling in rs31-1 mutant plants and plants constitutively overexpressing At-RS31 (Petrillo et al., 2014). By identifying At-RS31 direct targets through iCLIP and transcripts that are differentially alternatively spliced in response to altered At-RS31 levels, we gain insights into the downstream processes controlled by At-RS31.