Precision in RNA degradation: helicases, ribosome and RNA polymerase stalling, and RNA modifications

Abstract

Errors in transcription and RNA processing generate aberrant transcripts that can produce truncated, nonfunctional, or dominant-negative proteins. RNA surveillance pathways, centered on the RNA exosome, recognize diverse processing defects and initiate targeted RNA degradation. These mechanisms also regulate the short half-life of chromatin-associated noncoding RNAs through co- and/or post-transcriptional degradation. This review examines how the RNA exosome achieves substrate specificity, focusing on its interactions with the helicase cofactors mRNA transport 4 (MTR4) and superkiller 2 (SKIV2L) and the modulatory role of RNA modifications such as N⁶-methyladenosine. The broad spectrum of RNA exosome targets underscores its central functions in transcription, translation, genome integrity, and cell fate determination.

Translating RNA exosome structure to function

A central aim in the field of RNA degradation is to understand how functional and nonfunctional transcripts are programmed for degradation at the proper time and place following synthesis and completion of functionality. Without properly timed and targeted RNA degradation, cells can retain RNA transcripts, causing downstream problems such as activation of the inflammatory response, accumulation of toxic protein products, and DNA damage due to DNA–RNA hybrids [1,2]. The RNA exosome prevents the accumulation of incorrectly or prematurely processed transcripts from protein-coding genes, decreases the level of long noncoding RNAs (lncRNAs), and facilitates the turnover of nucleus-resident transcripts. The RNA exosome core is composed of nine catalytically inactive subunits (referred to as Exo9), possessing a barrel-like structural shape, with 3′ → 5′ exoribonuclease subunits DIS3/DIS3L positioned at the base of the structure and EXOSC10 at the top of the structure, thus forming the complete 11-subunit complex of the RNA exosome (or Exo11) [3,4]. An RNA helicase must be present to remodel RNA–protein complexes and unwind double-stranded nucleic acids before any substrate can pass through the narrow pore connecting Exo9 with DIS3/DIS3L. Two RNA helicases are primarily associated with the RNA exosome: MTR4 (Mtr4 in Scerevisiae) and SKIV2L (Ski2 in Saccharomyces cerevisiae) [5,6]. More recently, other helicases such as DDX1 and DDX60 have been shown to interact with the RNA exosome [7,8]. The RNA exosome is active in both the nucleus and cytoplasm, but the complex can utilize different catalytic subunits and partners with different helicases in each. In the nucleus, it functions primarily through the DIS3 RNase subunit and the MTR4 RNA helicase. In contrast, in the cytoplasm, the RNA exosome partners with the catalytic subunit DIS3L and the RNA helicase SKIV2L. Adding a further level of spatial complexity, DIS3 is actively excluded from the nucleolus to facilitate rRNA processing by EXOSC10 [9].

As just discussed, forming different complexes provides spatial complexity and substrate discrimination. At the same time, modifications to the RNA itself provide a level of temporal complexity, with dynamic base modifications signaling for degradation or protection at different points in the RNA life cycle.

This review explores and synthesizes many of the studies that have defined RNA exosome substrates and the structures of the complexes that deploy RNA exosome activity. We also speculate on potential mechanisms that these complexes employ to select their specific classes of substrates.

Nuclear RNA exosome helicase co-complexes

The nuclear RNA exosome depends on the helicase MTR4, which primarily acts within three major complexes: NEXT (Nuclear Exosome Targeting), TRAMP (Trf4/Air2/Mtr4p Polyadenylation), or PAXT (PolyA Exosome Targeting). To avoid confusion between different model systems and to focus on the biochemical activities of these complexes, we summarize nuclear RNA exosome cofactor complexes and RNA exosome subunits in Table 1.

Table 1.

RNA exosome cofactor complexes and subunits across evolution

RNA exosome cofactor complexes
Complex	S. cerevisiae	S. pombe	Mammalian	Primary Substrates
NEXT	Absent	Absent (MTREC-like)	MTR4•ZCCHC8•RBM7	Unstable ncRNAs: PROMPTs, eRNAs, CUTs, and uridine-rich 3′ tails [10-15]
PAXT	Absent	MTREC (Mmi1•Red1•Mtl1)	MTR4•ZFC3H1•PABPN1	Polyadenylated RNAs, unexported mRNAs, and snoRNA host genes [12,15,16]
TRAMP	Mtr4•Air1/2•Trf4/5	Mtr4•Air1/2•Cid14	hTRAMP (ZCCHC7•PAPD5/7•MTR4)	rRNA precursors, snoRNAs, snRNAs, and tRNAs [17]
SKI	Ski2•Ski3•Ski8	Ski2•Ski3•Rec14	SKIV2L•TTC37•WDR61	mRNAs with stalled ribosomes, cytoplasmic ncRNAs, and viral RNAs [5,18-20]
RNA exosome subunits
S. cerevisiae	S. Pombe	Human	Archaea	Role in Complex
Csl4	Csl4	EXOSC1	Csl4	Structural. Forms the cap of the barrel with EXOSC2 and EXOSC3. Channels RNA and assists with RNA binding through putative KH and S1 domains [4,21]
Rrp4	Rrp4	EXOSC2	Rrp4	Structural. Forms the cap of the barrel with EXOSC1 and EXOSC3. Channels RNA and assists with RNA binding through putative KH and S1 domains. [4]
Rrp40	Rrp40	EXOSC3	Absent	Structural. Forms the cap of the barrel with EXOSC1 and EXOSC2. Channels RNA and assists with RNA binding through putative KH and S1 domains. [4]
Rrp41	Rrp41	EXOSC4	Rrp41	Structural component of the Exo9 barrel in eukaryotes. Catalytic in archaeal RNA exosome complex. Binds to DIS3. [4,21]
Rrp46	Rrp46	EXOSC5	Absent	Structural component of the Exo9 barrel. Binds the EXOSC1-3 cap. [4]
Mtr3	Mtr3	EXOSC6	Absent	Structural component of the Exo9 barrel. Binds the EXOSC1-3 cap. [4]
Rrp42	Rrp42	EXOSC7	Rrp42	Structural component of the Exo9 barrel. Binds the EXOSC1–3 cap. [4]
Rrp43	Rrp43	EXOSC8	Absent	Structural component of the Exo9 barrel. Binds to DIS3. [4]
Rrp45	Rrp45	EXOSC9	Absent	Structural component of the Exo9 barrel. Binds to DIS3. [4]
Rrp6	Rrp6	EXOSC10	Absent	Catalytic subunit that binds adjacent to the Cap. Implicated in the RNA exosome’s activity in ribosome biogenesis and small RNA maturation. [3,9,11]
Rrp44	Rrp44	DIS3	Absent	Main catalytic subunit of the nuclear RNA exosome. Binds at the bottom of the barrel and cleaves RNA. [4,9]
Mpp6	Mpp6	MPP6	Absent	RNA Exosome cofactor that mediates Mtr4 binding to the RNA exosome core. [22-25]
Rpr47	Cti1	C1D	Absent	Contributes to bridging Mtr4 and the RNA exosome and interacts with Exosc10. [6,25]

Cryo-EM studies reveal that NEXT (MTR4, ZCCHC8, and RBM7) is organized around ZCCHC8, whose C-terminal domain regulates substrate access to the RNA exosome (Figure 1A) [10,26]. RNAs are encapsulated by MTR4 and translocated through its helicase channel via ZCCHC8-enhanced ATP hydrolysis [27]. The NEXT complex primarily degrades unstable non-polyadenylated RNAs such as enhancer RNAs (eRNAs), promoter upstream transcripts (PROMPTs), or cryptic unstable transcripts (CUTs) (Table 1) [11,12,28]. Biochemical experiments demonstrate that the NEXT complex most effectively unwinds substrates with long unstructured 3′ RNA ends, particularly when the tail is U-rich, but much less effectively when it is A-rich [27,29]. These experiments bring the questions of targeting and anti-targeting to the forefront of the RNA processing field. Because RNA is a lower diversity substrate than proteins and other molecules, understanding how stability is carefully monitored and controlled across several classes of transcripts is challenging. Understanding how NEXT selects its substrates can shed light on how the cell selects certain RNAs as coding or noncoding.

Figure 1. Nuclear RNA exosome partner complexes.

The nuclear RNA exosome depends on MTR4, which is weakly active on its own, but when its partner complex engages its substrate, it can bind to a scaffold that enhances its ATPase activity and, therefore, processivity. These complexes can then recruit the RNA exosome with EXOSC10 and DIS3 in the case of NEXT and PAXT, and EXOSC10 alone for TRAMP. (A) Unprocessed R-loops and unprocessed ncRNAs are largely recognized by RBM7, the RNA-binding protein of the NEXT complex [26]. ZCCHC8 is then able to recruit MTR4, and then the complex can recruit the nuclear Exo11 complex. These transcripts are then degraded [11]. (B) Polyadenylated R-loops and polyadenylated ncRNAs or unexported mRNAs are recognized by PABPN1, which binds to ZFC3H1 and recruits MTR4 and subsequently Exo11 for degradation [13]. (C) The TRAMP complex is a nucleolar complex. It contains PAPD for RNA binding and ZCCHC7 as a scaffold. RNA is not always fully threaded through the RNA exosome core but processed by EXOSC10 to yield appropriate rRNA and snoRNAs [17]. (D) Co-transcriptional activity of the RNA exosome occurs through backtracking at various stages of transcription. Steric factors such as negative and positive supercoiling can cause the 3′ end to be exposed, and the RNA exosome then removes the stalled polymerase and resolves the nascent transcript hybrid. This figure was created using BioRender.com. NEXT: Nuclear Exosome Targeting Complex; PAXT: Poly(A) tail Exosome Targeting; Pol II: RNA Polymerase II; TRAMP: Tfr4/Air2/Mtr4p Polyadenylation.

While NEXT processes aberrant ncRNAs, many RNAs that are polyadenylated can fail in later steps of RNA processing, especially during RNA export, accumulating and necessitating another layer of RNA quality control. Originally, MTR4 was identified in a screen searching for genes whose loss resulted in an accumulation of polyA+ foci in yeast, showing that this pathway also affects later steps in RNA quality control [30]. Later, it was shown that MTR4, within the PAXT complex [MTR4, ZFC3H1, and poly(A)− binding protein PABPN1], targets longer polyadenylated transcripts, particularly snoRNA host genes [12,13]. Unlike NEXT, which prefers uridine-rich 3′ ends, PABPN1 confers affinity for poly(A)+ substrates. Within the PAXT complex, ZFC3H1 likely acts as a scaffold, recruiting factors such as ARS2 for early termination, analogous to the Schizosaccharomyces pombe MTREC complex (Figure 1B) [13,15]. A recent study has shown that ZFC3H1 is loaded onto the 5′ end of pre-mRNAs and that shorter, single-exon genes are more susceptible to RNA exosome-mediated degradation, partially because ZFC3H1 is closer to the 3′ end of the pre-mRNA [16]. This mechanism is similar to the HUSH (Human Silencing Hub) complex heterochromatinizing intronless LINEs (Long Interspersed Nuclear Elements) to prevent aberrant transcription [31].

In addition to the quality control aspects of NEXT/PAXT, the RNA exosome and MTR4 are also crucial for some routine cellular functions. The TRAMP complex, which was originally characterized in S. cerevisiae (Mtr4, RNA-binding proteins Air1/2, and noncanonical polymerase Trf4/5), is less characterized in humans but facilitates rRNA maturation and the 3′ processing of snoRNAs, snRNAs, and tRNAs [17]. Yeast lack a NEXT complex, with its functions being partially overtaken by the TRAMP complex [14,15]. There are human homologs of all the TRAMP complex subunits, but the human TRAMP complex (Figure 1C) (hTRAMP) is not well characterized. hTRAMP is localized to the nucleolus, suggesting that it plays a role in rRNA processing or maturation [32]. The hTRAMP subunit ZCCHC7 is overexpressed in DLBCL tumors, and this overexpression causes a defect in pre-rRNA processing, specifically pre-5.8S rRNA. This defect also results in a shift toward a proteome that closely resembles lymphoma proteomes [33]. Overexpression of ZCCHC7 can result in a defect in functional ribosome formation, resulting in a downregulation of proteins whose translation is relatively more sensitive to ribosome stoichiometry.

MTR4 exhibits significantly enhanced activity within the NEXT and TRAMP complexes in duplex unwinding assays. MTR4 was first demonstrated to have helicase activity on RNA–RNA hybrids and DNA–RNA hybrids in vitro [34]. More recent studies have shown that MTR4’s helicase and ATPase activities are weakly coupled, and in the case of S. pombe, Mtr4 is completely decoupled and unable to unwind RNA on its own. Components of the TRAMP complex coordinate interdomain interactions and allow duplex unwinding [35]. Scaffold components of the NEXT and TRAMP complexes, particularly ZCCHC8 and ZCCHC7, are able to link these domains in MTR4 and RNA-binding proteins such as RBM7 [26]. RBM7 and RNA-binding proteins increase duplex unwinding by both tethering the complex to the substrate during unwinding and/or increasing the rate of initiation [26,36]. RBM7’s preference for uridine-rich sequences can explain NEXT’s selectivity for unprocessed substrates such as PROMPTs. These studies illustrate how cofactors such as RNA-binding proteins and scaffolds confer substrate specificity upon MTR4, thereby directing RNA exosome activity.

In addition to controlling RNA exosome substrate specificity through modulating Mtr4 activity in different complexes, NEXT, TRAMP, and PAXT show different subnuclear localizations [32,37-39]. The TRAMP complex primarily localizes to the nucleolus, the NEXT across the nucleoplasm, and PAXT to polyA+ foci. Controlling subnuclear localization adds a second layer of specificity to MTR4–RNA exosome targeting, protecting transcripts within a particular compartment from other complexes, as well as enhancing the activity of complexes through increasing local concentration. RNA structure itself can drive phase changes that could help sort MTR4 complexes or regulate their formation [40,41]. While much work has yet to be done, a model of nuclear RNA substrate specificity for RNA exosome-mediated degradation is emerging, where MTR4 is weakly active on its own, but spatially regulated complex formation stimulates unwinding. It is not clear if this increase is due to increased affinity/residence time for its proper substrate, an increase in processivity, or both. In the case of the S. cerevisiae TRAMP complex, Mtr4’s initiation rate is increased, but its helicase activity is unchanged [36]. In the case of the S. pombe TRAMP complex, Mtr4 has ATPase activity but requires cofactors to coordinate interdomain interaction to unwind substrates, indicating that helicase activity itself does not change due to cofactor binding [35]. In contrast, during NEXT complex stimulation, ATPase activity is increased, suggesting that helicase activity might also be increased [26,27]. Further experiments are required to conclusively differentiate between these hypotheses.

MTR4’s nucleic acid hybrid unwinding activity was recognized through its role in supporting the development of adaptive immunity in vitro and in vivo [34]. MTR4 supports immunoglobulin diversification during class-switch recombination (CSR). Additionally, the NEXT complex and its cofactor MPP6 physically interact with AID in B cells [42]. During CSR, the enzyme AID acts on ssDNA created through transcription over G-rich repetitive sequences called switch regions, leaving three-stranded structures called R-loops. The looped-out DNA strand is an excellent target for AID activity [43,44]. To generate mutational symmetry, these R-loops must be resolved. MTR4 and RNA exosome subunit mutant cells exhibit defective mutational symmetry during AID activation, reflecting that they are required for efficiently resolving nucleic acid hybrids at the Ig loci [34].

Despite the progress in understanding MTR4’s role in adaptive immunity, many outstanding questions regarding MTR4 remain. The precise complex within which MTR4 acts during CSR and other stages of B-cell development remains unknown. Immunoglobulin class switching requires both spatial specificity and temporal specificity by the RNA exosome, so understanding what cofactor complexes are acting is a crucial step toward understanding the exact mechanism of RNA processing in adaptive immunity. It is still unclear how the RNA exosome is recruited to the immunoglobulin switch regions or VDJ segments, especially since these loci are only active for an exquisitely limited period and are targeted only to very specific genomic loci to allow for optimal AID activity. MTR4 must act very specifically to resolve the R-loops at switch regions, but paradoxically, it must not act so quickly as to resolve the structure before AID acts on the single-stranded non-template strand. This cannot be fully explained by sequence specificity, because dramatically divergent sequences are still able to go through class-switch recombination [43,44]. Likely, there exists some mechanism to greatly stimulate MTR4 activity in response to polymerase stalling or DNA damage, and understanding this mechanism is a promising approach to exploring RNA exosome targeting.

Beyond CSR, RNA surveillance is a common motif across B-cell development. In early B-cell development, deficiencies in RNA helicases, as well as nuclear RNA exosome components such as DIS3, result in a direct defect in VDJ recombination that can be partially rescued with prearranged alleles [45,46]. It is unclear through which mechanisms the RNA exosome and other cofactors are acting in this context. Furthermore, DIS3 mutations are one of the most common driver mutations in multiple myeloma, but how DIS3 mutations are tolerated so well in plasma cells and how DIS3 mutations contribute to the etiology or progression of multiple myeloma is only beginning to be resolved [47,48].

MTR4’s role in B cells opens several avenues of research and raises important questions regarding its enzymatic property of resolving DNA–RNA and RNA–RNA hybrids in vivo. Most importantly, it involves the resolution of secondary DNA structures such as R-loops, G-quadruplexes, and stem loops, all of which are dependent upon the activity of the RNA exosome complex. While RNA exosome activity in preserving transcription regulation through the maintenance of genomic integrity and epigenetic regulation was originally shown in class-switching B cells and ES cells [11,28], recent studies have expanded this mechanism to other cell types, including germ cells, neurons, plasma B cells, and the context of repetitive DNA element control [47,49-51]. Supporting these observations, it has also been demonstrated that an imbalance in RNA exosome stoichiometry, through the overexpression of a single subunit, increases genomic instability in S. cerevisiae [52]. Indeed, the nuclear activity of the RNA exosome, through its core helicase MTR4, is now recognized as a central effector in coupling transcription control with genomic integrity. How MTR4 complexes with the RNA exosome to identify specific RNA substrates present at specific loci in the genome will remain a driving question for the field.

Co-transcriptional RNA surveillance in resolving RNA polymerase II backtracking

Following escape from promoter-proximal pausing, RNA polymerase II (Pol II) frequently undergoes elongation-phase pausing or stalling due to diverse impediments such as DNA secondary structures, DNA damage, chromatin modifications, and higher-order chromosomal architecture, all of which can cause other termination pathways to fail [53]. Pol II pausing/stalling promotes the formation of stable DNA–RNA hybrids (R-loops) through prolonged hybridization of the nascent RNA transcript with the DNA template strand. The buildup of negative supercoiling behind and positive supercoiling ahead of the elongating polymerase creates a favorable DNA topology for DNA–RNA hybrid formation, particularly in the presence of single-strand nicks in the non-template DNA strand or during Pol II backtracking (Figure 1D). These stalled RNA polymerases create genomic instability through collisions with the DNA replication complex [54,55].

In these contexts, RNA exosome-mediated 3′ end trimming of the nascent transcript has been proposed to facilitate the resolution of stalled Pol II complexes [28,56]. These foundational studies were performed in the context of activation-induced cytidine deaminase (AID)-mediated somatic hypermutation, a process tightly coupled to transcriptional pausing and stalling during elongation [57]. Parallel to RNA exosome activity, the Integrator complex cleaves nascent RNA, and the protein phosphatase 2 complex removes RNA Pol II [58]. It is not clear whether the RNA exosome and Integrator complexes act redundantly or concordantly in the process, but overlapping sets of transcripts are sensitive to both complexes [59]. Recent studies have identified N-MYC as a critical cofactor that potentiates RNA exosome function in resolving Pol II backtracking, thereby limiting co-transcriptional DNA–RNA hybrid accumulation and suppressing aberrant innate immune activation triggered by persistent R-loops [60,61]. Collectively, these studies implicate the helicase activity of the RNA exosome cofactor MTR4 as important for the efficient resolution of stalled Pol II complexes. However, the mechanistic basis by which the MTR4 complex accesses the nascent RNA embedded within a DNA–RNA hybrid at a backtracked polymerase remains unclear, particularly as the stalled polymerase protects RNA 3′ ends. Additionally, through which RNA exosome co-complexes MTR4 acts and during which stage in transcription remain unclear. Addressing this fundamental question will require creative high-resolution structural and biochemical studies to elucidate how RNA surveillance machinery is physically coupled to transcription elongation and genome stability maintenance.

Cytoplasmic RNA exosome-associated helicases

Cytoplasmic RNA degradation and quality control pose a fundamentally different challenge for the cell compared with nuclear RNA degradation. Unlike the nucleus, which has a much wider variety of cryptic transcripts, the cytoplasm has a relatively smaller fraction of aberrant/malformed RNA. Additionally, mRNAs in the cytoplasm need to be preserved for translation without being subjected to promiscuous degradation. Several specialized pathways exist to eliminate specific misprocessed RNAs, such as those with retained introns and early stop codons [62]. Additionally, small RNAs, such as microRNAs, add a layer of complexity to RNA processing in the cytoplasm, which is not limited to RNA degradation [63]. In this section, we specifically limit our focus to the cytoplasmic RNA exosome.

Like the nuclear RNA exosome, the cytoplasmic RNA exosome also depends on helicase activity, primarily the helicase SKIV2L. Similar to the nuclear exosome cofactor complexes, SKIV2L partners with the scaffolding protein TTC37 (SKI3) and the protein WDR61 (SKI8) to enable its helicase activity. Unlike the nuclear RNA exosome, the SKI complex is the sole identified cofactor complex for the cytoplasmic RNA exosome [64]. Understanding substrate selectivity is challenging in this context because RNA must be protected from degradation for a longer temporal period to allow for multiple rounds of translation.

The SKI complex binds to both ribosome-covered and ribosome-free regions of mRNA and assists the cytoplasmic RNA exosome in degrading long 3′UTR-containing mRNAs, poorly translated mRNAs, and cytoplasmic lncRNAs [18,65]. SKIV2L recruits the assembled complex to regions occupied by active ribosomes [66]. Once the SKI complex is localized to 80S ribosomal complexes, it extracts mRNA in the 3′ → 5′ direction. Facilitating efficient mRNA degradation by the SKI complex requires a buffer of at least 19 nucleotides from the ribosome’s peptidyl (P) site [67].

HBS1L3 (hSKI7) acts as a bridging factor between the ribosome and the RNA exosome by directly binding to the ribosome and recruiting the RNA exosome complex. As a result, there is a continuous path of RNA threading from the 80S ribosome unit through the SKIV2L helicase [68]. In its default resting state, SKI is in a closed conformation. The 3′–5′ RNA-unwinding channel of SKIV2L is blocked by a gatekeeping module composed of the N-terminal domain of SKIV2L [5,19]. In response to ribosome stalling or termination events, the SKI complex can efficiently extract mRNA 3′ ends from stalled ribosomes in an ATP-dependent manner and undergo a conformational switch to an active, open state. In this state, the SKIV2L gatekeeping module detaches from the SKIV2L helicase domain, which allows the RNA 3′ end to exit the helicase and enter the RNA exosome ribonuclease channel [19,69].

Structural insights reveal the precision and coordination of co-translational mRNA degradation through the SKI complex. Aberrant mRNA associated with collided or stalled ribosomes is recognized and channeled from the ribosome, through the SKI complex, and into the RNA exosome for specific degradation. This autoinhibitory action is crucial for limiting RNA exosome activity in the cytoplasm to poorly translated or defective transcripts, providing quality control without preventing efficient translation [19,68].

Ribosome stalling resolution through the SKI complex

Through the SKI complex, the cytoplasmic RNA exosome is intimately coupled to resolving stalled ribosomes. However, multiple pathways exist, and their interdependence has not been fully characterized. Because the SKI complex is associated with degrading mRNAs on stalled ribosomes, its functions are often coupled with the action of ribosome-recycling complexes. One such complex is the ribosome-resolving complex of PELO and its partner HBS1L, which plays a central yet context-dependent role in distinct mRNA degradation pathways, often facilitated by the ATPase ABCE1 (Figure 2) [69]. During no-go degradation (NGD), the primary trigger for mRNA degradation is ribosome stalling, which can lead to ribosome collisions. This typically occurs when ribosomes encounter impediments such as stable mRNA secondary structures [70,71]. Within NGD, PELO resolves ribosomes stalled at the 3′ end of mRNA after an initial cleavage event, binding to the ribosomal A site once it becomes empty. The downstream outcome for the mRNA is endonucleolytic cleavage and degradation (Figure 2) [74].

Figure 2. Specificity of mRNA degradation during ribosome collision and stalling.

In NGD, the primary trigger for mRNA degradation is ribosome slowing or stalling due to mRNA secondary structures, rare codons, or damaged nucleotides. The yeast Dom34–HBS1 complex recognizes these collided ribosomes and recruits Cue2 for endonucleolytic cleavage of these RNAs [70,71]. In the case of NSD, the absence of a stop codon leads to ribosomes translating into the mRNA’s poly(A) tail and results in the synthesis of polylysine tracts. This can also cause the ribosome to stall or collide in the poly(A) tail region of the mRNA. As such, PELO can recognize these stalled ribosomes and recruit the SKI complex as an upstream factor that results in the recruitment of the RNA exosome. The mRNA is processed in a 3′ → 5′ RNA exosome-mediated degradation [72,73]. This figure was created using BioRender.com. NGD: no-go degradation; NSD: nonstop degradation.

Conversely, in nonstop degradation (NSD), mRNA degradation is primarily triggered by the absence of a stop codon, leading to ribosomes translating into the mRNA’s poly(A) tail and generating polylysine tracts. The SKI complex acts as a key upstream factor in this pathway. Here, PELO’s specific role is to resolve ribosomes stalled on these poly(A) tails after they have skipped the stop codon, again binding to the empty A site. The downstream outcome for the mRNA in NSD is typically 3′ → 5′ RNA exosome-mediated degradation, which is facilitated by the bridging action of the SKI complex [72,73]. In both NGD and NSD, the actions of PELO, HBS1L, and ABCE1 are critical for initiating the precise removal of problematic mRNA–ribosome complexes, contributing significantly to mRNA quality control (MQC) (Figure 2).

Recent studies unveiled a novel synthetic lethal interaction between the PELO–HBS1L and SKI complexes within the MQC pathway, highlighting a critical vulnerability in specific cancer subtypes [75,76]. Notably, HBS1L and HBS1L3 are splicing variants of the same gene, with HBS1L primarily acting in the PELO complex and HBS1L3 serving as a crucial bridging factor within the SKI complex [76].

In distinct genetic contexts, including 9p21.3-deleted tumors (where biallelic deletion of the FOCAD gene leads to SKI destabilization) and microsatellite instability-high (MSI-H) tumors with mutations in TTC37, a critical SKI component, the destabilization of the SKI complex significantly impairs the efficient extraction of mRNA from stalled ribosomes [76].

This compromised state renders cancer cells uniquely dependent on PELO for their viability. When this pathway is disrupted, particularly through SKI complex deletions, PELO emerges as a critical factor in maintaining the specificity of identifying and resolving these problematic transcripts. The catastrophic consequence of losing both components—SKI deficiency combined with PELO depletion—leads to a severe accumulation of stalled ribosomes and a dramatic induction of the unfolded protein response (UPR), consistently upregulating gene expression programs driven by NF-κB, p53, and IRE1α activation [75]. This outcome demonstrates a profound failure in the cell’s precise MQC mechanisms. Consequently, PELO and HBS1L have been identified as promising novel therapeutic targets.

The question arises as to how the cytoplasmic RNA exosome complex, with SKIV2L or other RNA helicase-containing complexes, recognizes substrate RNAs for degradation. As evident from previous sections, there is no single mechanism for cytoplasmic RNA targeting to the RNA exosome, but rather several interdependent pathways.

N⁶-methyladenosine modification and its readers stimulate RNA degradation in the nucleus and cytoplasm

We have focused on protein factors that target RNA substrates for degradation, but now we turn to substrate modifications that alter RNA stability. Epitranscriptomics, the study of RNA modifications, is a mature field, with several modifications (as well as their coordination) having been characterized. We refer readers to the following reviews for a more thorough examination of this subject [77-79]. Here, we will focus on one of the most prevalent modifications, N⁶-methyladenosine (m^6A). m⁶A is a reversible RNA modification that arises from the coordinated activity of methyltransferases (m⁶A writers), especially METTL3/METTL14, and demethylases (m⁶A erasers), especially ALKBH5 and FTO. These marks have the potential to be recognized by m⁶A-recognizing reader proteins. m⁶A reader proteins contain a YTH (YT521-B homology) domain, which can specifically recognize the m⁶A mark and bind to the modified base. Upon binding, they often act as a scaffold to recruit downstream functions. YTH family proteins are not homologous outside of their YTH domain. Each can effectuate different processes in response to m⁶A methylation, indicating that m⁶A marks can be highly context dependent and that the regulatory grammar of epitranscriptomics will shed light on the transcriptional and regulatory life cycle of RNAs. Other proposed models argue that YTHDF1/2/3 are largely redundant proteins that all serve to enhance the degradation of cytosolic RNAs and that effects on translation might have been due to secondary effects [80]. Furthermore, IGFBP proteins have also been identified as m⁶A readers that stabilize RNAs through binding to m⁶A, representing a potential parallel pathway for regulating m⁶A transcripts [81]. Some reported functions of m⁶A reader proteins are listed in Table 2.

Table 2.

Summary of m⁶A reader proteins and their downstream functions

Reader protein	Localization	Key function(s)
YTHDF1	Cytoplasm	Promotes translation initiation, stress granule formation/sorting, binds CCR4–NOT [82,83]
YTHDF2	Cytoplasm	Recruits degradation machinery: CCR4–NOT (deadenylation), UPF1 (decapping), and HRSP12/RNase P/MRP (cleavage) [84,85]
YTHDF3	Cytoplasm	Recruits PAN2-PAN3 deadenylase complex; synergizes with YTHDF2 to degrade transcripts and with YTHDF1 to promote translation [80,86]
YTHDC1	Nucleus	Recognizes m6A on nuclear RNAs; recruits NEXT complex and RNA exosome for degradation [42]
YTHDC2	Nucleus/Cytoplasm	Has intrinsic helicase activity; promotes translation of m6A-modified mRNAs; plays a role in meiosis [87,88]
IGF2BP	Cytoplasm	Promotes m6A-containing RNA stability [81]

The METTL3/METTL14 heterodimer catalyzes m⁶A modification of RNA exosome-substrate intronic RNAs and drives their subsequent degradation. Degradation of transcription start site-associated RNAs (TSS-RNAs), eRNAs, and long intergenic ncRNAs is partly m⁶A modification dependent, as these RNAs are increased in Mettl3-deficient B cells [42]. SμGLT is an RNA exosome-sensitive lncRNA that accumulates within Mettl3-depleted cells, as the transcript is protected from 3′ end processing and RNA exosome-mediated degradation [42]. It has been directly shown that RNA exosome-sensitive substrates are m^6A-modified, particularly eRNAs, TSS-RNAs, and LINE elements [89-92]. m⁶A deposition is suppressed by the exon-junction complex, leading to spliced RNAs being protected from m⁶A [89].

The regulation of LINE1 retrotransposons extends beyond epigenetic DNA modifications and includes post-transcriptional control via RNA modifications. Specifically, LINE1 RNA transcripts are established targets of m⁶A modification, which plays a critical role in both LINE1 RNA degradation and broader transcription control [90]. In this pathway, the nuclear m⁶A reader YTHDC1 directly recognizes m⁶A-modified LINE1 RNA. This recognition facilitates the recruitment of the NEXT complex and the nuclear RNA exosome, leading to the rapid degradation of LINE1 RNA in the nucleus [51,93]. Ultimately, this pathway ensures the tight control of LINE1 transcript levels.

Further highlighting the importance of LINE1 transcription control is the dynamic interplay between m⁶A, LINE1 RNA, and chromatin states. In the absence of the m⁶A demethylase FTO, LINE1 RNA exhibits increased m⁶A levels [92]. This, in turn, leads to a reduction in LINE1 RNA abundance, likely due to the m⁶A–YTHDC1–NEXT complex degradation pathway. Importantly, this decrease in LINE1 RNA impacts the chromatin landscape, promoting a more closed chromatin state at LINE1-containing genomic regions. Conversely, restoring LINE1 RNA levels can reestablish an open chromatin state.

During RNA degradation, a reversible mark on RNA could be used to designate RNA for degradation or further processing. Indeed, m⁶A is often deposited on a wide range of RNAs and is removed during processing steps. Recent studies have identified m⁶A modification of RNA exosome-substrate intronic RNAs and their subsequent degradation [94]. These results position m⁶A as a central mark in distinguishing mRNAs from other RNA species (Figure 3).

Figure 3. Nuclear and cytoplasmic RNA modification-mediated degradation.

During B cell class-switch recombination, the m⁶A modification in the lncRNA SμGLT is recognized by the nuclear m⁶A reader YTHDC1. This recognition leads to the recruitment of the nuclear adaptor MPP6 and the RNA exosome for 3′ → 5′ exonucleolytic degradation [42]. In TSS-associated LINE1 RNAs, m⁶A is also recognized by YTHDC1; however, it now recruits the NEXT complex and the RNA exosome for LINE1 RNA degradation [93]. In the cytoplasm, m⁶A is recognized by the cytoplasmic m⁶A-readers YTHDF2 and YTHDF3, which process mRNA for degradation through the CNOT1/CCR4 pathway or the PAN2/PAN3 pathway, respectively [84]. In addition, the presence of both m¹A and m⁶A in a transcript can result in an interaction between HRSP12 and YTHDF2 that leads to the recruitment of RNase P/MRP endonuclease, resulting in endonucleolytic cleavage [95]. This figure was created using BioRender.com. LINE1: Long Interspersed Nuclear Element 1; m¹A: 1-Methyladenosine; m⁶A: N6-methyladenosine; m⁷G: 7-Methylguanosine; NEXT: Nuclear Exosome Targeting; SμGLT: Switch-μ Germline Transcript; TSS: transcription start site.

Convergence of helicases, ribosome stalling, or RNA modification to support substrate recognition and degradation by the RNA exosome

The findings discussed earlier support a unified view of RNA surveillance, in which RNA exosome targeting and activity on various substrate RNAs in the nucleus or the cytoplasm are not dictated by a single mark or pathway but emerge from overlapping networks of chemical, mechanical, and enzymatic activities. While helicases isolate RNAs from DNA–RNA hybrid configurations or other secondary structures and provide the directionality for subsequent RNA exosome activity, ribosomes provide real-time functional assessment, and RNA modifications provide contextual signals. Understanding how these layers are coordinated is essential for elucidating how cells maintain transcriptome integrity and how failures in RNA surveillance contribute to cancer, neurodegeneration, and viral pathogenesis. This integrated framework also raises the possibility that selective modulation of helicase recruitment or RNA modification readers could be exploited to therapeutically redirect RNA fate.

As mentioned earlier, particularly regarding m⁶A, chemical modification of RNA has the potential to provide specificity toward ribosome behavior and helicase accessibility. m⁶A can locally destabilize RNA secondary structure and recruit YTH domain reader proteins, such as YTHDF2, which promotes RNA degradation through interactions with the deadenylation and degradation machinery (Table 2). On the other hand, m⁶A also alters ribosome dwell time, increasing susceptibility to NGD or nonsense-mediated decay [96]. Terminal uridylation, catalyzed by TUTases, marks RNAs—particularly deadenylated mRNAs and small RNAs—for rapid degradation, including RNA exosome-dependent pathways [97]. Conversely, noncanonical polyadenylation by TRAMP in the nucleus creates short, unstructured tails that function as degradation-promoting handles rather than stabilizing elements [22].

Beyond m⁶A, defective or stress-induced modifications, such as oxidized bases, hypomodified tRNAs, or aberrant rRNA modifications, can directly induce ribosome stalling or impair RNP assembly, thereby indirectly promoting RNA exosome targeting [98-100]. These observations support a model in which RNA modifications tune the kinetic balance between ribosome stalling during translation and RNA degradation. Further discovery of other mechanisms such as localization of RNAs in existing or transiently formed membraneless organelles, new types of secondary structures, and unknown small biomolecules that function as secondary messengers will provide a finer understanding of RNA degradation targeting to functional and nonfunctional RNAs in the nucleus and the cytoplasm.

Pathogenic variants of the RNA exosome

Missense mutations in genes encoding structural subunits of the RNA exosome give rise to a class of rare diseases referred to as exosomopathies. Large deletions or loss-of-function mutations are not observed because some level of residual RNA exosome activity is required for viability. Several studies have characterized RNA exosome subunit mutations with diverse clinical presentations. For example, EXOSC3 mutations are associated with pontocerebellar hypoplasia type 1b (PCH1b), a severe neurological disorder defined by atrophy of the cerebellum and pons. Mutations in EXOSC8 and EXOSC9 cause the spinal muscular/motor disorders PCH1c and PCH1d, respectively, whereas EXOSC2 pathogenic variants cause mild or borderline cerebellar atrophy and SHRF syndrome (short stature, hearing loss, retinitis pigmentosa, and distinctive facies) [101]. EXOSC5 mutations also result in neurological defects. The reported EXOSC5 mutations have been shown to affect their association with other RNA exosome subunits [102]. It is unclear whether clinical phenotypes resulting from RNA exosome mutations that affect complex stability are the result of a lower rate of complex formation, formation of a less functional complex, or formation of a defective subcomplex.

SKI complex mutations cause trichohepatoenteric syndrome, a genetic disease that leads to primary immunodeficiency [103]. Until recently, the cause of the immunodeficiency was unclear, but it was recently discovered that these mutations result in defects in the earlier stages of B-cell development during VDJ recombination and that VDJ recombination depends on RNA exosome activity [45,46]. The exact mechanisms or cofactors through which the RNA exosome might act during VDJ recombination remain unclear. In the case of SKI mutations, a defect in the cytoplasmic RNA exosome may alter the distribution of RNA exosome complexes, resulting in a shift toward cytoplasmic localization and limiting RNA exosome availability to the Ig loci.

Mutations, translocations, and copy number gains in the hTRAMP complex scaffold ZCCHC7 are observed in high-grade double-hit lymphoma. Recreating these mutations in lymphoma cell lines results in an accumulation of 5.8S + 40 pre-rRNA, leading to a rewired proteome [33]. This could represent a pro-adaptive response by lymphoma cells to manage the increased translational burden, either by increasing the rate of ribosome biogenesis or by being better prepared to handle a burst of translation. A similar mechanism of altered ribosome biogenesis might exist in DIS3 mutant multiple myeloma [104]. Although DIS3 is excluded from the nucleolus, mutant DIS3 with lower processivity could lead to a similar increase in pre-rRNA species if the RNA exosome core is prevented from cycling to the nucleolus for extended periods. This could explain why DIS3 mutant multiple myeloma cases might partially resist proteasome inhibitor-based therapies [105].

Concluding remarks

All newly generated RNAs are regulated by complex processing and assembly pathways to equip them for diverse cellular functions. Defects arise in RNA transcripts and RNA–protein complexes because of this large-scale oversight over an extensive pool of transcripts. Although several quality control mechanisms exist, this review focuses on the RNA exosome, a versatile multisubunit complex integral to RNA degradation, maturation, and surveillance. The RNA exosome mainly targets ncRNA through cofactor complexes, but while RNA exosome substrates have been well defined, the precise temporal and spatial targeting dynamics are poorly understood. One attractive targeting mechanism is RNA modification. RNA modification can mark RNAs for processing (or as processed) or for degradation, providing temporal control of when RNAs are processed. This might involve recruiting machinery such as the RNA exosome in the case of m⁶A modifications. RNA modifications could be incorporated by default and removed as part of normal processing, similar to how some RNA modifications or DNA methylation contribute to self versus nonself recognition in innate immunity (see Outstanding questions).

Outstanding questions.

How do RNA exosome cofactors contribute to specificity toward the recognition of specific types of RNAs in the nucleus and cytoplasm?

In the nucleus, through what mechanisms and with which cofactors does the RNA exosome recognize nascently made RNAs produced at different stages/states of RNA polymerase II transcription?

How efficiently does the SKI complex recognize stalled ribosomes in vivo? How do RNA modifications contribute to RNA substrate sorting?

What, if any, sequence features determine whether noncoding RNAs are targeted for degradation through the RNA exosome?

As our understanding of the precise functions of RNA exosome subunits becomes clearer, we will be able to make definite conclusions about their unique roles in disease and their potential as therapeutic targets.

Highlights.

Compartmentalized helicase–exosome complexes confer spatiotemporal specificity in nuclear and cytoplasmic RNA decay, preventing toxic transcript accumulation.

Conformational gating ensures 3′ → 5′ RNA degradation is selectively activated by ribosome stalling, safeguarding functional mRNAs.

Epitranscriptomic marks act as central signals for decay, recruiting exosome adaptors and directing nuclear/cytoplasmic surveillance of noncoding RNAs.

Exosomopathies and cancer vulnerabilities arise from mutations in exosome cofactors, exposing therapeutic targets.