Introduction

Living organisms are continuously exposed to a barrage of environmental and internal stimuli, which can interfere with mRNA-coding capacity of the DNA. To ensure proper gene expression, mRNAs are routinely inspected for errors before making deleterious proteins. Decoding genetic information into functional molecules is essential to cellular life. Living organisms have a sophisticated and tightly coordinated gene expression system that ensures a variety of physiological functions and maintains genome integrity. Upon internal or external environmental challenges, the gene-expression system can interfere with the mRNA-coding capacity of DNA. Therefore, to ensure proper gene expression, mRNAs are routinely inspected for errors before deleterious proteins with negative effects are produced. One such mechanism, which regulates gene expression at the post-transcriptional level, is nonsense-mediated mRNA decay (NMD), which prevents the production of truncated non-functional and/or aberrant dominant-negative proteins by scavenging faulty mRNAs containing premature termination codons (PTC) as well as physiologically inappropriate transcripts (Maquat and Carmichael 2001; Lewis et al. 2003; Green et al. 2003; Mendell et al. 2004; Weischenfeldt et al. 2008; Wengrod et al. 2013; Hug et al. 2016). NMD prevents the production of truncated non-functional and/or dominant-negative proteins by scavenging aberrant mRNAs containing premature termination codons (PTCs). The NMD-triggering features, such as PTC, are caused by mutations in genes (such as nonsense or frame-shift mutations), genomic rearrangements, RNA splicing errors and alternative translation initiation. In addition, normal mRNAs that are targeted by NMD can be products of alternative splicing (AS) that introduce PTC (Kalyna et al. 2012; Drechsel et al. 2013) or contain cis-elements that recognize their normal termination codon as premature (Kervestin and Jacobson 2012; Popp and Maquat 2013; Schweingruber et al. 2013; Popp and Maquat 2014; Peccarelli and Kebaara 2014). As a result, NMD, as a central mechanism, is pivotal to gene expression, biochemical reactions, hormonal signalling, the circadian clock, reproduction, and gene evolution, which support cellular functions (Raxwal and Riha 2023).

In this review, we concisely discuss a basic outline of the molecular mechanism and impact of NMD in eukaryotes. Known distinctive features of plant NMD are comprehensively explained along with an overview of a possible unified network of stress-related genes and the effects of NMD on other cellular activities. Overall, this article describes the NMD mechanism in plants under different stress conditions and shows how manipulations of certain NMD factors can provide information on the less explored interlinkage of RNA metabolism, stress response, and adaptation in plants. A breakthrough in plant breeding can be achieved by harnessing accumulating knowledge of NMD’s natural pathways and their dynamic regulation by environmental and developmental signals to improve abiotic resistance and enhance the defence response against pathogens.

Nonsense-mediated mRNA decay pathway in eukaryotes

NMD is a translation-coupled mRNA quality control pathway that selectively eliminates mRNAs with PTCs or NMD-triggering features. NMD is now used in a variety of ways by eukaryotes to destroy PTC-containing transcripts. The NMD core machinery contains three trans-acting up-frameshift proteins (UPF)—UPF1, UPF2, and UPF3—which have a conserved regulatory role in coordinating the NMD pathway across eukaryotic species (Pulak and Anderson 1993; He et al. 1997; Lykke-Andersen et al. 2000; Mendell et al. 2000; Wang et al. 2001; Serin et al. 2001; Gatfield et al. 2003; Kerényi et al. 2008). UPF1 (a superfamily I RNA helicase) contains both ATP-dependent 5′-to-3′ helicase activity and RNA-dependent ATPase activity (Yoine et al. 2006; Chakrabarti et al. 2011) that binds directly to the PTCs harboring mRNA during translation, and interacts with the translation termination complex. However, UPF1 also binds non-specifically to 3′ untranslated regions (UTR) of both NMD-sensitive and NMD-insensitive mRNAs (Hogg and Goff 2010; Hurt et al. 2013; Kurosaki et al. 2014; Lee et al. 2015). Activation of NMD involves UPF1/SMG2, an SF1-family ATP-dependent RNA helicase that undergoes cycles of phosphorylation and dephosphorylation (Yoine et al. 2006; Dai et al. 2016), but the underlying mechanism needs to be further studied.

Loss of UPF1 has varied effects across species. For instance, UPF1 in mice is crucial for embryonic development but impairing the NMD pathway does not affect viability in Caenorhabditis elegans (Hodgkin et al. 1989). Eukaryotic NMD mechanism can be divided into three broad steps: recognition of target mRNA, commitment for mRNA decay, and degradation of NMD targets (which may occur inside processing bodies (P-bodies)) (Fig. 1).

Fig. 1
figure 1

Outline of steps involved in the mechanism of eukaryotic nonsense-mediated decay (NMD). a After splicing, the mRNAs retain some of the NMD-substrate signatures, including upstream (u) ORF, premature termination codon (PTC), intron retention (IR) (predominantly in plants), and a long 3′ UTR. b NMD is triggered and recognized by eRF1 and eRF2 complexes present on the translating ribosomes during any round of translation. The ribosome stalls on the mRNA at the PTC and cap factors on the 5′ end of the mRNA recruit SMG1–UPF1 onto the exon-junction complex (EJC). The resultant complex is called SURF (SMG1–UPF1–eRF1–eRF3). SMG1 phosphorylates UPF1 and the resulting conformational changes cause ribosomes, eRF, and polypeptide chains to dissociate. c SMG5/7 complex dephosphorylates UPF1 via PP2A,releasing SMG1. d mRNA enters P-bodies where degradation proceeds via decapping and deadenylation complexes. After mRNA degradation, NMD factors are recycled back into the cytoplasm for another round of NMD

Full size image

The other two UPFs (UPF2 and UPF3) have been reported to stimulate UPF1 activities (helicase and ATPase). Evidence suggests that NMD acts through multiple or branched pathways that require different sets of other NMD factors, most of which are undiscovered (Bühler et al. 2006). The recent consensus is that UPF1 is the main factor for NMD activation (Yoine et al. 2006; Dai et al. 2016), but the underlying mechanism needs further study.

Recognition of NMD targets

NMD targets have distinctive characteristics such as mRNAs with PTCs or NMD-triggering features. In mammalian cells, PTC situated more than 50–55 nucleotides upstream of the last 3′ exon–exon junctions in mRNA can elicit NMD. This condition is referred to as the 50–55 nucleotide rule (Nagy and Maquat 1998). However, the definition of a nonsense stop codon as pre-mature premature and the position of PTCs from the exon-exon junction that triggers NMD vary across organisms.

Another criterion for NMD targets is the presence of an upstream open reading frame (uORF), which is a primary regulatory element of gene expression at the translation or post-transcription level and lies upstream of the main ORF in 18–57% of plant mRNAs (Wang et al. 2023); specifically, only > 50 amino acid-long uORFs have been documented to elicit effective NMD (Nyikó et al. 2009; Saul et al. 2009). However, Cymerman et al. (2022) revealed that 70 amino acid-long uORF-harboring transcripts can escape NMD in plants, indicating that the conditions that expose uORF to NMD are different for mammalian and plant cells. Other factors that have been recognized to trigger NMD in vertebrates include the presence of 3′ untranslated region (UTR)-located introns, isoforms generated by AS, and a UGA codon for selenocysteine (Schweingruber et al. 2013). Furthermore, among several factors contributing to NMD efficiency, the 3′ UTR length is also crucial (Hogg and Goff 2010; Hug et al. 2016), with a length of ≥ 350 nucleotides usually triggering NMD (Kalyna et al. 2012; Kertesz et al. 2006). However, Raxwal et al. (2020) showed that 70% of the NMD-targeted transcripts have PTC and only 30% are caused by unusually long 3′ UTR, indicating that most transcripts that undergo NMD contain PTC. Although the criteria for PTC recognition are not yet clear, defining the substrate for NMD in plants requires more robust research. In addition, the efficiency of the plant NMD response also depends on PTC position and cis-elements (introns and 3′ UTR), similar to the mammalian and yeast systems (Chiba and Green 2009; Sureshkumar et al. 2016; Ohtani and Wachter 2019). However, studies have shown that, unlike mammals, intron retention in the 3′ UTR is pivotal for PTC-recognition events in plants (Hori and Watanabe 2007). Apart from the sequence-specific features, complexes that are recruited and remain attached to the mRNAs throughout their processing, such as the exon junction complex (EJC)—a group of proteins involved in target recognition and surveillance processes—also dictate recognition of NMD targets (Kertesz et al. 2006; Nyikó et al. 2013). However, EJC have a controversial role in the initiation of NMD (Kurosaki et al. 2018). Based on the EJC’s involvement in initiating target recognition, two different mechanisms are considered, namely 3′ UTR EJC-dependent NMD and 3′ UTR EJC-independent NMD.

Untranslated region EJC-dependent NMD

The EJC factors are promiscuously recruited approximately 20–24 nucleotides upstream of exon–exon junctions of nearly 80% of the mRNA during splicing (Le Hir et al. 2000, 2001; Lejeune et al. 2003; Singh et al. 2012). This complex consists of conserved core members, including eukaryotic translation-initiation factor 4 A3 (eIF4A3), MAGOH, and Y14 proteins (Mabin et al. 2018), while UAP56, REF, ALY, and RNA-binding protein with serine-rich domain 1 (RNPS1) are found attached to the periphery of the EJC core at subsequent stages of the NMD pathway (Fig. 2). Studies have shown that the dynamic composition of the EJC plays a crucial role in facilitating mRNA processes such as splicing (Ashton-Beaucage et al. 2010; Roignant and Treisman 2010; Wang et al. 2014), nuclear export (Schmidt et al. 2006; Le Hir et al. 2001), and translation (Diem et al. 2007; Isken et al. 2008; Chazal et al. 2013). During translation, the EJC and proteins associated with the 5′ UTR or coding region of the mRNA are removed by the translocating ribosomes. In the case of normal mRNAs, the ribosome removes all EJC during translocation up to the stop codon of the last exon. Furthermore, the ribosomes do not transcribe the 3′ UTR and therefore, the proteins associated with the 3′ UTR of the transcript remain intact (Dostie and Dreyfuss 2002; Sato and Maquat 2009). The translation-termination process is then initiated by the ribosome at the stop codon interacting with the poly (A)-bound proteins at the 3′ mRNA. When the ribosome approaches the termination codon, a ternary termination complex consisting of eRF1, eRF3, and GTP is recruited to the A site of the ribosome (Joazeiro 2017). The eRF3 hydrolyses GTP, causing conformational changes in eRF1. Consequently, the polypeptide is hydrolysed from the polypeptidyl-tRNA at the processing site and the nascent peptide is released.

Fig. 2
figure 2

Schematic representation of exon-junction complex (EJC) core components in plant nonsense-mediated mRNA decay (NMD). During spliceosome assembly, various EJC factors are associated with splicing intermediates. The core EJC factors include UPF1, UPF2, UPF3, Y14, MLNS1, SAP18, ACINI RNPS1, Ref/Aly, PININ, eIF4A3, and MAGOH and the transient factors (Tap, UAP56, SRm160, p15, and PYM) are situated on the periphery of the EJC core

Full size image

When the PTC are present upstream of the stop codon, the translocating ribosomes stall at the PTC, leaving downstream EJC intact on the mRNAs (Popp and Maquat 2018). The distance of the stalled ribosome to the 3′ poly (A) tail may be large enough to initiate termination. Consequently, the release of the ribosome from the mRNA is delayed, thus providing enough time for the recruitment of other NMD factors onto the EJC (Bühler et al. 2006). The NMD factors UPF3X–UPF2 are recruited onto the 3′ EJC downstream of the PTC, forming an effective 3′ UTR EJC (Kim et al. 2001; Chamieh et al. 2008; Buchwald et al. 2010). At the PTC, the translation-termination complex recruits UPF1 and suppressor with morphogenetic effect on genitalia (SMG1), forming a ternary complex termed SMG1–UPF1–eRF(SURF) (Kashima et al. 2006). The SURF complex interacts with the downstream 3′ UTR EJC, and further forms a decay-inducing complex. Now the 3′ UTR recruits UPF1–SMG1 from the termination complex to the EJC (Chamieh et al. 2008; Chakrabarti et al. 2011). Another RNA helicase, DEAH box polypeptide 34 (DHX34), facilitates the interaction of UPF1 with UPF2 of the EJC (Hug and Caceres 2014), and helicase activity of UPF1 is then stimulated by UPF2, followed by UPF1 phosphorylation at the C terminus by SMG1 (Yamashita et al. 2001; Kashima et al. 2006; Kurosaki et al. 2014). This event commits the mRNAs to degradation (Kashima et al. 2006). However, the work of Karousis et al. (2020) challenges the concept of NMD activation by stalled ribosomes at the PTC of NMD-sensitive mRNAs. Toeprint assay on human cell lysate in vitro revealed similar ribosomal density at the PTC and a normal termination codon in NMD-sensitive and NMD-insensitive mRNAs, respectively, indicating that ribosome stalling at the PTC is not exclusively related to NMD activation.

3′ Untranslated region EJC-independent or 3′ UTR-dependent NMD

The mRNAs without 3′ UTR are also found to be targeted by the NMD pathway and these mRNAs are documented to contain cis-elements like introns and unusually long 3′ UTR, that recognize normal termination codon (NTC) on the mRNA as premature (Buhler et al. 2006; Singh et al. 2008; Matsuda et al. 2007; Wang et al. 2002; Zhang et al. 1998), however, the 3′ UTR-dependent NMD mechanism is poorly understood as compared to the 3′ UTR EJC- dependent NMD pathway. It has also been shown that mRNAs lacking the 3′ UTR EJC are also targeted by the NMD pathway. Such mRNAs have elements such as introns and an unusually long 3′ UTR, which causes the recognition of normal termination codons as premature (Bühler et al. 2006; Singh et al. 2008; Matsuda et al. 2007; Wang et al. 2002; Zhang et al. 1998). However, the 3′ UTR-dependent NMD mechanism is less understood than the 3′ UTR EJC-dependent NMD pathway. Interestingly, 3′ UTR EJC-independent NMD is found in some endogenous mRNAs without PTC but with an > 1 kb long 3′ UTR in mammalian systems (Singh et al. 2008). Transcriptome-wide profiling of human cells indicates the presence of UPF1 on mRNAs that undergo 3′ UTREJC-dependent NMD, indicating that UPF1 binds non-specifically to most mRNAs, but dissociates from the non-target mRNAs (Kurosaki et al. 2014; Lee et al. 2015; Imamachi et al. 2017). Presumably, the probability of UPF1 binding to the long 3′ UTR increases due to the presence of a wide space on the mRNA (Imamachi et al. 2017). Studies have shown that UPF1 generally binds with the secondary structure, forming G-rich sequences (Hurt et al. 2013) on the mRNA; these structures impede the translocation of UPF1 and thus probably increase the chance of UPF1 phosphorylation, followed by subsequent steps of NMD. 3′ UTREJC-independent NMD has been documented during pioneer and later rounds of translation, unlike 3′ UTREJC-dependent NMD, which is found to be active predominantly during the pioneer round (Hosoda et al. 2005; Kim et al. 2005; Choe et al. 2014).

Commitment to mRNA decay

NMD is a translation-coupled process in which messenger ribonucleoproteins are exported to the cytoplasm and efficiently accessed by ribosomes (Trcek et al. 2013; Halstead et al. 2015; Popp and Maquat 2015). Any defect in the translational machinery has been reported as a failure of the NMD process (Chiba and Green 2009). In plants, the targeted mRNAs associated with another NMD factor, PYM, are exported from the nucleus into the cytoplasm through the nuclear pore complex (Xu and Chua 2011). 3′ UTR EJC-mediated dependent NMD largely takes place during the initial round of translation in the cytoplasm (Maquat et al. 2010). The cap-binding complex (CBC) CBP20–CBP80 heterodimer remains attached to the mRNA and helps recruit ribosomes, which are replaced by eIF4E in a later round of translation (Choe et al. 2012). However, the degradation of mRNA is so rapid that most mRNAs are probably already degraded in CBC-bound form, before the replacement of CBC from the 5′ end of the mRNA by eIF4E (Kugler et al. 1995; Cheng and Maquat 1993; Belgrader et al. 1994). In Saccharomyces cerevisiae, in addition to 3′UTR EJC-dependent NMD, the presence of 3′ UTR-mediated NMD ensures the degradation of CBC and eIF4E-bound mRNAs (Hosoda et al. 2005). The cohort of mRNAs in CBC-bound form (which are degraded only after the initial round of translation) and that of mRNAs bound to eIF4E (which are degraded in later rounds of translation) have an equal probability of being degraded, but relative amounts are unknown. Hoek et al. (2019) characterized certain mRNA attributes that cause its degradation by NMD in any round of translation: the number and distance of introns both downstream and upstream of the PTC, and the exon sequence downstream of the PTC. After recognition of the termination codon, determined by either 3′ UTR EJC-dependent or 3′ UTR-mediated NMD, the SURF complex activates SMG. SMG1 phosphorylates the serine- and threonine-rich regions of UPF1 (Kerenyi et al. 2013), which is the interlinking factor between ribosomal translation and the NMD pathway, and phosphorylation of UPF1 is a mandatory step for initiation of the mRNA-decay process (Franks et al. 2010). In a phosphorylated state, UPF1 also interacts with the EJC-bound UPF (UPF2 and UPF3). Structural analysis has revealed that UPF2 serves as an adapter connecting UPF1 to the EJC core and mRNA (Chamieh et al. 2008).

Phosphorylated UPF1 has a greater affinity for dephosphorylation elements (Dai et al. 2016). As a consequence, dephosphorylation factors, including SMG5/7–SMG6 heterodimer complexes, are recruited to the phosphorylated UPF1. In mammals, each SMG has a distinct role; SMG6 cleaves mRNA, whereas both SMG5 and SMG7 recruit deadenylase and the decapping machinery (Okada-Katsuhata et al. 2012). SMG5/7 and SMG6 interact with phosphatase 2A proteins and promote the dephosphorylation of UPF1 (Ohnishi et al. 2003). However, the timing of the association between phosphatase 2A and SMG5/7 is still unclear. Overall, phosphorylation of UPF1 causes the translational machinery to dissociate while dephosphorylation induces EJC dissociation and initiation of decay processes. The dephosphorylated UPF1 state is reported to be favoured for entry into P-bodies, where further degradation steps are performed by decapping and deadenylation complexes. Thus, the mechanistic coupling of translating ribosomes with other proteins deposited at the EJC drives NMD under strict regulation in a sequential manner.

Localization of mRNA into P-bodies for decay

The P-bodies are dedicated mRNA-degrading areas in the cytoplasm that contain decay factors. However, some other models speculate that P-bodies serve as a storage region for untranslated mRNA under stress conditions. While NMD can take place in P-bodies in plants, it does not strictly occur in P-bodies in mammalian cells (Stalder and Mühlemann 2009; Mérai et al. 2013). Therefore, a more balanced description of the intracellular site of NMD in different organisms is warranted. Targeted mRNAs are localized into P-bodies by two concomitant processes—mRNA degradation and ribosomal recycling (Schuller et al. 2018). The late phase of the NMD process, which occurs in the cytoplasm, has been reported to differ in plants and mammals. Interestingly, the CBC is essential in mammals whereas in plants, the role of the CBC in the later NMD phase is not necessary (Dzikiewicz-Krawczyk et al. 2008; Mérai et al. 2013; Dai et al. 2016). During the late phase, the N- and C-terminal-phosphorylated UPF1 conjugates with the NMD complex and SMG7, for localization of the whole complex from the cytoplasm into the P-bodies in the plants (Okada-Katsuhata et al. 2012; Mérai et al. 2013). Phosphorylation of UPF1 is a mandatory step for initiation of the mRNA-decay process (Lejeune et al. 2003; Kurosaki et al. 2014; Isken et al. 2008; Franks et al. 2010). This step also checks further translation rounds (Isken et al. 2008). In plants, degradation of the targeted mRNA can be carried out using two different pathways, depending on the presence or absence of plant ortholog of SMG7. In the presence of SMG7, the targeted mRNA enters the SMG7–XRN4 5′-to-3′ decay pathway whereasin its absence, it follows the UPF1–XRN4 5′-to-3′ decay pathway. The XRN family has been reported to contain prominent decapping enzymes (Chiba and Green 2009)in plants, including cytoplasmic XRN4. Mérai et al. (2013) showed that in the plant NMD process, degradation is predominantly through the latter decay pathway which includes two important sequential steps: translational repression and decapping of the target mRNA. The mRNAs with PTC are translationally repressed before decay (Maquat and Carmichael 2001).

In the presence of SMG7, the phosphorylated UPF1 activates the SMG7–XRN4 pathway. This pathway is responsible for the localization of mRNAs into P-bodies and their further degradation in a 5′-to- 3′ manner (Dai et al. 2016). In contrast, in the absence of SMG7, XRN4 has also been found to activate 5′-to-3′ decay. The decay is further accompanied by 3′-to-5′ decay in the presence of poly(A)-binding protein and eRF1, indicating that SMG7 is dispensable for transcript degradation via NMD. Before degradation, the 5′ and 3′ ends of the target mRNAs are respectively decapped and deadenylated in the NMD process (Kim and Maquat 2019). The direction of decay is under debate, i.e., whether it starts with deadenylation or decapping. However, co-immunoprecipitation studies have shown that the decapping machinery loads onto mRNA earlier than the deadenylase units, speculating that decapping occurs before the deadenylation of NMD-targeted mRNAs (Kurosaki et al. 2019).

Furthermore, decapping may occur before entry into the P-bodies, due to SMG7 recruitment or as a result of UPF1 dephosphorylation on the mRNAs (Mérai et al. 2013). The homolog of XRN1 (yeast and metazoan) from the decapping enzyme family, acts on the 5′ end and proceeds to catalyse uncapped mRNA degradation. XRN acts on both polyadenylated and deadenylated mRNAs. However, RNA degradome analysis of xrn4 mutants demonstrated the presence of a higher amount of decapped, deadenylated transcripts than the polyadenylated mRNA (Nagarajan et al. 2019). This indicates that XRN4 prefers deadenylated mRNAs over polyadenylated mRNAs. In mammalian NMD, the SMG6–SMG7 complex is responsible for the recruitment of the decapping and deadenylation machinery (Ohnishi et al. 2003; Okada-Katsuhata et al. 2012). However, in plants, despite the lack of metazoan SMG6 ortholog, XRN4 is still recruited and plays an important role in the decay (Xu and Chua 2011). This indicates that XRN4 does not depend absolutely on SMG6 in plants but is regulated by multiple degradation pathways. The Arabidopsis genome, for instance, encodes several PIN/PIN-like domain-containing endoribonucleases that generate 5′ monophosphate on 3′ fragments, which are then degraded by XRN4 without the involvement of SMG6 (Nagarajan et al. 2019). Studies have shown that uridylation of NMD-targeted mRNAs presumably also stimulates degradation (Kurosaki et al. 2019). But again, despite its crucial role, NMD and its underlying molecular mechanisms remain poorly understood.

Ribosome recycling to replenish components of the translational machinery

Ribosome recycling represents a key process in translational control to generate a rapid response under normal physiological conditions (Celik et al. 2015). Ribosomal recycling occurs by splitting of the ribosome subunits, mediated by ATP-binding cassette E1 protein, (ABCE1). In this process, the 40S and 60S subunits of ribosomes are split by recruiting conserved ATP binding cassette E1 protein (ABCE1), an iron-sulphur (Fe-S) cluster-containing protein that terminates the ribosome with eRF1 and initiates ribosomal recycling. Depending upon the nature of the termination codons—normal versus premature, ribosome recycling occurs through two distinct mechanisms (Pisarev et al. 2010). In normal mRNAs, the normal termination codon is located near the last exon, and during translation, the ribosomes remove all upstream EJC. Once the normal termination codon is reached, eRF1 and eRF3 associate with the terminating ribosome and facilitate peptide and eRF3 release. In this scenario, eRF1 recruits ABCE1, at the site previously occupied by eRF3. After the release of the peptides, ABCE1 dissociates the ribosomal subunits and enables ribosomes to undergo multiple rounds of translation (Hellen 2018). For mRNAs with PTC, the ribosome associated with eRF1 and eRF3 stalls at the PTC upstream of the EJC, and another assembly, the UPF1–SMG1 complex, is recruited at the EJC situated nearest to the PTC. After releasing and dissociating the peptide and eRF3, respectively, the ABCE1 factor is prevented from recruitment onto the previously occupied eRF3 site of the ribosome due to the interaction of eRF1 and UPF1–SMG1 (Celik et al. 2015), preventing multiple rounds of translation. Moreover, a few reports have suggested that ribosome recycling in the case of mRNAs harbouring PTC is mediated by the ATPase activity of UPF1 binding to the extended 3′ UTR (Neu‐Yilik et al. 2017). Notably, the role of ABCE1 in PTCs containing transcripts cannot be ignored as a defect in ABCE1 has been reported with an in efficient NMD pathway. The role of ABCE1 in PTC-containing transcripts cannot be ignored, because it is pivotal in translation initiation and termination. Moreover, a decrease in ABCE1 is anticipated to affect the biogenesis of functional proteins, potentially leading to the suppression of NMD (Annibaldis et al. 2020). This evidence indicates a possible role of ABCE1 in ribosome recycling of both normal and nonsense mRNAs (Zhu et al. 2020). In the case of 3′ UTR dependent NMD depleting ABCE1, the ribosome profile shows high ribosome occupancy at 3′ UTR. The ribosome profile shows high ribosome occupancy at 3′ UTR with decreased activity of ABCE1, in the case of 3′ UTR-dependent NMD. This may result in increased codon read through, which displaces UPF1 and EJC from the 3′ UTR and inhibits NMD. The roles of ABCE1 in ribosomal recycling during translation are critical and warrant further study for a more in-depth understanding of the NMD mechanism.

Nonsense-mediated mRNA decay as a fitness-enhancement strategy in plants

Several research studies have shown that NMD is not only involved in stress responses in animals but also in plants. The NMD-mediated gene response is an integral process of molecular regulation to maintain cellular homeostasis and trigger defence responses in plants to environmental and/or internal stresses (Jeong et al. 2011; Raxwal and Riha (2016) and illustrated in Fig. 3. Several studies have shown that plants demonstrate defence responses against environmental and internal stresses through NMD-mediated gene responses (Raxwal and Riha 2016; Jeong et al. 2011) (Fig. 3). In addition to transcripts associated with NMD triggering features, normal physiological mRNAs are also targeted by NMD, indicating the role of NMD in shaping cellular transcriptome by regulating the abundance of normal mRNAs (Lewis et al. 2003; Green et al. 2003; Mendell et al. 2004; Weischenfeldt et al. 2008; Wengrod et al. 2013). However, the extent of the NMD response to a specific stimulus has yet to be estimated. Nonsense-mediated mRNA decay has been found to play a dominant role during development, for instance, in the early auxin response during shoot generation in Arabidopsis thaliana (Ohtani and Wachter 2019). In upf1 and/or upf3-1 mutant plants, NMD controls multiple auxin-related genes in association with the post-transcriptional regulatory system to distribute auxin in the callus (Ohtani and Wachter 2019). The upf-mutant plants have also been reported to influence the early auxin-response genes and the transcription factors cup-shaped cotyledon 1 and 2 (Ohtani and Wachter 2019). In that report, auxin signalling—this governs the transition of the callus from a root to shoot-like state in the shoot-inducing medium—was found to function abnormally in NMD mutants, indicating the role of functional NMD in development. Moreover, AtSF1, similar to the SF1 splicing factor in mammals, was found to be active in AS and to regulate ABA sensitivity during seed germination (Jang et al. 2014), and Cruz et al. (2014) showed that exogenous ABA application on A. thaliana seedlings changes the AS pattern in relation to the serine–arginine (SR) protein family (Cruz et al. 2014). This may be due to the presence of phosphatase HYPERSENSITIVE TO ABA1 that binds and dephosphorylates SNF1-RELATED PROTEIN KINASE 2 (SnRK2) kinases, especially SnRK2.6 (OPEN STOMATA 1), and inhibits the transduction of ABA signal (Sybilska and Daszkowska-Golec 2023). Similarly, EBF2contains a long 3′ UTR that binds to ETHYLENE INSENSITIVE2 (ethylene signal transducer), and induces ethylene triggering in an upf2 mutant in A.thaliana (Raxwal and Riha 2023). Nonsense-mediated mRNA decay has been documented to be downregulated in both mammals and plants under stress conditions to elevate expression levels of the defence-response genes. The plants can identify and respond to invading pathogens (biotic stress exposure), further it is suggested that plants possess specific reactivity and memory similar to that found in adaptive immune system in mammals. Unlike vertebrates, the immune system of plants is not adaptive, but plants are capable of launching very specific, self-tolerant immune responses in the form of resistance proteins that identify and respond to invading pathogens (Spoel and Dong 2012). Factors of NMD are responsible for the routine surveillance of gene-expression profiles and associated intracellular toxicity in plants (Park et al. 2020). In the pioneer round of translation, some PTC-containing mRNAs escape degradation and are translated to misfolded, truncated polypeptides. The production of misfolded polypeptides has also been reported from messenger ribonucleoproteins undergoing ribosomal stalling (Meriin et al. 2018). Accumulation of these misfolded polypeptides causes proteotoxic stress, which leads to cellular apoptosis. Under normal circumstances, the cell engages in the removal of these misfolded polypeptides via different mechanisms, including refolding of the truncated polypeptides by molecular chaperones and elimination through the ubiquitin–proteasome system. In fact, this system has been found to be under UPF1 regulation via its E3 ubiquitin ligase activity (Feng et al. 2017). However, when these mechanisms are overwhelmed or otherwise insufficient, another mechanism—the CTIF–eEFIAI–DCTNI pathway of aberrant polypeptide elimination—is activated (Park et al. 2020). This pathway is also under the control of hyperphosphorylated UPF1 through aggresome formation.

Fig. 3
figure 3

Steps in nonsense-mediated mRNA decay (NMD) under stress conditions such as pathogen attack, drought, osmotic stress, or temperature fluctuations in plants. Components of P-body regulate plant stress-related genes. RNA cap-binding protein heterodimer CPB80–CBP20 targets mRNA that contains PTC. EJC components are formed by splicing factors (RNPS1, UAP56, SRM160, PININ), exporting factors (UAP56, REF/Aly, TAP, Y14, or MAGOH), and a group of four proteins forming the core of the EJC (MAGOH, Y14, eIF4A3, and MLN51). Various stresses induce protein kinases (MEK4, MKK4/5, MPK3/6, SnRK2, MAP4K) to phosphorylate, decapping complexes, which in turn mediate mRNA decay via XRN4 (5′-to-3′ degradation)

Full size image

This evidence indicates a direct link between NMD factors (such as UPF1) and pathways involved in eliminating faulty polypeptides (Park et al. 2020). Hyperphosphorylated UPF1 facilitates the retrograde transport of truncated misfolded polypeptides into aggresomes for elimination (Durand et al. 2016). As the misfolded polypeptides decrease to an ambient level, intracellular proteomic stress is also managed in the cell. The factor UPF1 ensures the proper quality and quantity of mRNAs and polypeptides. Therefore, UPF1 is implicated not only in mRNA-quality control but also in protein surveillance as part of post-translational events.

Coupled AS–NMD pathways for maintenance of cellular homeostasis and adaptation to stress

Approximately 95% of the pre-mRNAs of higher eukaryotes undergo AS, which drives protein diversity, mRNA stability, and translational efficiency (Hamid and Makeyev 2014). Importantly, the generation of AS variants is tightly regulated and highly responsive to internal and external cues (Staiger and Brown 2013). However, the AS process is also prone to ineffective intron removal, leading to either intron- or exon-derived PTC (Neumann et al. 2020). Therefore, NMD coupled with AS provides an efficient strategy for systematic molecular regulation of gene expression (Neumann et al. 2020). The former is crucial because most of the mRNA targets are critical AS-NMD variants (Kalyna et al. 2012) (Table 1). In tomato genotypes T250 and T270 grown under water deficit, low nitrate, or a combination of these conditions, the number of splice events was higher than in their non-stressed counterparts, suggesting that AS might regulate transcription (Ruggiero et al. 2022). Biological and abiotic stresses constantly threaten plant survival and affect their growth and development. Consequently, stress-response genes tend to undergo AS to prepare them for unfavourable conditions (Szakonyi and Duque 2018; Laloum et al. 2018). However, there is limited evidence correlating the number of splice products with protein variations under stress conditions in plants. Desai et al. (2020) analysed chromatin immunoprecipitation (ChiP) data sets in humans and mice and suggested that NMD couples with AS via a poorly understood mechanism termed regulated unproductive splicing and translation (RUST). This mechanism has also been documented in yeast. The isoform products generated through this unproductive AS process have been reported in heterogeneous nuclear ribonucleoproteins, splicing factors, and genes encoding Serine/arginine rich (SR) proteins which function as splicing enhancers. Most of the isoform transcripts generated through RUST contain PTC, primarily due to splicing errors that are subsequently targeted by NMD, thereby markedly decreasing the number of productive transcripts and reportedly reducing gene expression. It is well documented that RNA-binding proteins and splicing factors control their own expression levels via a negative feedback loop mediated by AS–NMD, where an excessive amount of RNA-binding protein binds its own pre-mRNA, resulting in AS and further inducing a PTC (Pervouchine et al. 2019). In plants, however, NMD regulation of genes encoding splicing factors through RUST remains to be investigated.

Table 1 Alternatively spliced (AS) variants of corresponding genes targeted by NMD in Arabidopsis thaliana
Full size table

Among the different types of AS, the intron-retention (IR) type has been majorly documented in plants (Chaudhary et al. 2019). In addition, the frequency of IR ranges from 28 to 64% in plants under specific developmental cues (Mandadi and Scholthof 2015). Although the actual reason for the occurrence of a high rate of IR is still obscure, its role in plants under various conditions, such as stress and development, is quite evident. Intron retention has been documented to be responsible for the production of most of the nonsense mRNAs in plants (Hori and Watanabe 2007). Some fractions of these nonsense mRNAs are degraded by NMD, while those that escape degradation form truncated proteins that contribute to gene expression in response to the plant's specific physiological status (Chaudhary et al. 2019). Under different stress conditions, certain factors (external and internal) are responsible for the modulation of the diverse pattern of the IR process that is directly correlated with a different set of proteins to confer the acclimatization process. It has been recently proposed that under stress conditions, plants decrease the translation of functional genes via AS and are primarily involved in producing transcript isoforms of stress-related genes to combat stress (Chaudhary et al. 2019). The presence of PTC in the first and last introns is due to selection, and not mere chance. It offers additional physiological ways to conserve energy that would otherwise be expended in the production of an excess number of proteins. Therefore, the coordinated coupling of AS with NMD induces a rapid stress-specific response. Alternative splicing often determines the total ratio of sense-to-nonsense transcripts produced under stress conditions. The same strategy has also been revealed in the functioning of circadian clock processes (Sureshkumar et al. 2016). The AS–NMD process has been identified as a prominent strategy to cope with toxicity mediated by biotic and abiotic stress conditions. The splicing ratio and its correlation with stress signals have yet to be elucidated in plants.

Abiotic constraints and NMD in plants

Abiotic stress requires efficient and timely responses by plants involving AS–NMD-mediated co- and post-transcriptional gene reprogramming, which regulates the level of stress-response factors and produces protein isoforms with specific properties (Staiger and Brown 2013; Filichkin et al. 2018). Recently, Shi et al. (2019) showed the potential role of AS–NMD in maca (Lepidiummeyenii) plants for high-altitude adaptation, where stressed plants produce erroneous transcripts via AS, which are then degraded by NMD. Moreover, the study confirmed the involvement of NMD-triggering elements, including SR proteins and long non-coding RNA in differential AS at high altitudes. Therefore, high-altitude adaptations might involve a cumulative response to AS and NMD processes via a complex pathway. Another significant finding includes the type of AS under stress. Previously, IR had been designated as a major type of AS occurring under stress conditions (Hori and Watanabe 2007). However, the alternative 3′ splice site type of AS was found to be prominent in maca plants for altitude adaptation. These variations in plants related to the type of AS under different stress conditions warrant further study.

Temperature is well-known to have a significant impact on the plant’s physiology and immune system. A change in temperature is often accompanied by changes in other abiotic factors, including light and humidity (Hua 2013). The temperature-dependent flowering response in Arabidopsis is carried out by FLOWERING LOCUS M (FLM). FLM represses another gene, FLOWERING LOCUS T (FT), which induces flowering. FLM mRNAs encode flowering-repression proteins and are subject to AS. At low temperatures, FLM primarily produces FLM β isoforms that form functional FLM proteins (Balasubramanian et al. 2006) which subsequently repress FT, and the floral transition is inhibited. At high temperatures, SR protein splicing regulators are differentially expressed with respect to temperature (Balasubramanian et al. 2006). These SR proteins modulate extensive splicing of FLM, and various splice variants (a major splice fraction containing PTC or NMD-triggering features) are produced at the expense of FLM β (Balasubramanian et al. 2006; Sureshkumar et al. 2016). FLM expression is downregulated by AS–NMD (Sureshkumar et al. 2016; Staiger and Brown 2013). This results in enhanced FT expression and causes flowering in Arabidopsis under high-temperature conditions (Sureshkumar et al. 2016; Balasubramanian et al. 2006). Interestingly, NMD mutants show more sensitivity to light than wild-type plants. The NMD-compromised Arabidopsis plants exposed to long days (16 h light) exhibited stunted growth and phenotypic abnormalities; short days (10 h light) produced no significant changes. In addition, NMD mutants under light stress are more vulnerable to other stresses. Moreover, NMD mutants that have been physically stressed or infected with a virus produce high levels of stress-related chemicals such as jasmonic acid and salicylic acid after long-day exposure. In another report, expression of the leucine-rich repeat (LRR) family was found to be modulated by NMD in a temperature-dependent manner. Showed that the expression level of three genes of the LRR family—AT1G72910, AT1G72940, and ADR1-LIKE 2—is altered by the NMD defence system under low-temperature conditions; upf1and upf3 double mutants have a high level of aberrant transcripts encoded by LRR-encoded genes at 16 °C. Thus, NMD regulates genes at lower temperatures which are crucial for NMD-mediated regulation of defence signalling in a temperature-dependent manner.

Nonsense-mediated mRNA decay to counteract pathogen invasion

Pathogen infection in crop plants accounts for nearly a quarter of all crop losses (Garcia et al. 2014) and new strategies are required to increase plant’s tolerance to pathogens. In addition to morphological and physiological defects, NMD-compromised plants can show visibly improved resistance to pathogen-mediated infections; for instance, upf1-5 and upf3-1 mutants in A. thaliana showed remarkable resistance to Pst DC3000 (Jeong et al. 2011). Yet, a few plant RNA viruses have been reported with long 3′ UTR (similar to NMD-targeted mRNA) which makes the viral genetic material (RNA) susceptible to NMD in plants (Cesari 2018). UPF1 targets RNA viruses and can reduce virus accumulation in most plants (Cesari 2018). Therefore, the RNA quality-control mechanism—NMD—of the host cells can be effectively used in responses against viral infections (Balistreri et al. 2017). Positive-sense single-stranded RNA ((+) ssRNA) is targeted by NMD, but there are no similar reports for negative-sense (−) ssRNA viruses. Tran et al. (2021) reported that the (−) ssRNA of influenza A virus does not elicit an effective NMD response in the host, despite having NMD-eliciting molecular patterns. In fact, the template of (−) ssRNA viruses, which act as precursors for new viral RNAs, is speculated to scarcely undergo translation, thereby bypassing the NMD-activation step and remaining intact for further infection (Tran et al. 2021).

Viruses and hosts co-evolve to infect and defend, respectively. Viruses such as Pea enation mosaic virus 2 (PEMV2), Rous sarcoma virus (RSV), and Turnip crinkle virus (TCV) have evolved strategies to circumvent NMD (Balistreri et al. 2017). RSV and TCV have NMD-resistant sequences. PEMV2 encodes a viral protein that inhibits NMD (May et al. 2020). Another strategy is the binding of viral elements to NMD factors to inhibit their functioning (Mocquet et al. 2015; Fiorini et al. 2018). Hepatitis C virus counteracts NMD by inhibiting the EJC recycling factor WITHIN BGCN HOMOLOG. May et al. (2018) reported that some viruses incorporate a pyrimidine-rich sequence before the gag stop codon and unstructured regions at the start of the 3′ UTR as an efficient strategy to bypass host NMD. Some viruses have also been reported to reprogram host NMD factors to aid in their replication process (Tran et al. 2021). Ajamian et al. (2015) documented the involvement of UPF1 in regulating viral protein synthesis as well as nuclear export through ATPase activity during viral infection. The helicase activity of UPF1 can also be reprogrammed and exploited to increase the infection efficiency of the newly synthesized virus particle. DNA viruses can also be affected by NMD. All viruses produce mRNAs and are therefore potential targets for host UPF1 (van Gent et al. 2021). As these studies are restricted to in-vitro analyses and provide only limited information on the actual NMD-mediated virus-counteracting mechanism, this area needs more attention (Tran et al. 2021). Genome recoding has recently been touted as a promising field for studying viral infection in crops (Kumar and Singh 2021). A comprehensive understanding of the pathogenesis and counterattack mechanism of plant viruses that evade host NMD will open new avenues for developing effective therapeutic approaches to rescuing agricultural crops from viral attacks.

Concluding remarks and future perspectives

Progress in plant NMD research has shifted the focus on the NMD pathway from a mere surveillance mechanism to a sophisticated stress-management strategy that confers accurate gene expression and cellular homeostasis. The NMD pathway has been well-studied in yeast, mouse, and human cells, but an in-depth understanding of plant NMD under stress conditions is still lacking. The compiled work proposes experimentally documented links between stresses (biotic and abiotic) and NMD pathways in plants that could provide a foundation for future investigations (Fig. 4). The important gaps that have been identified in our understanding of the NMD mechanism include undiscovered NMD-triggering cues, the degradation mechanism of truncated proteins, and the pathway by which NMD core factors identify the balance of sense and nonsense mRNA transcripts, and how NMD confers defence under multiple stress conditions. In addition, the mechanistic details of NMD factors in other metabolic processes remain to be worked out.

Fig. 4
figure 4

Nonsense-mediated mRNA decay (NMD) as a stress-management strategy in plants. Schematic illustration showing the downstream processes coordinated by NMD to revive cellular homeostasis throughout proteotoxic, biotic, and abiotic stresses in plants

Full size image

To understand NMD-mediated plant immunity, an investigation into the status of NMD factors during infection stages, and how NMD shapes the transcripts of plant receptors could reveal the underlying mechanism of NMD in plants for gene manipulation to improve plant’s stress tolerance. Crop diseases caused by viruses may be treated by targeting the mechanisms used by viruses to avoid ribosome frameshifting-triggered NMD (May et al. 2018). Moreover, understanding the interactions between plant development and defence responses and the NMD mechanism will assist in better understanding developmental shifts and eventually, metabolism alterations that support one event at the cost of another. Robust studies on NMD-associated stress-induced regulation of certain genes that are involved in plant defence and their downstream use would increase spatiotemporal knowledge of the underlying molecular mechanism, and ultimately provide directions for the development of strategies to optimize plant fitness at the species-specific level.