# Chapter 35. Eukaryotic mRNA Decay: Deadenylation, Decapping, NMD, and Regulated Pathways

## Scope Note

This chapter explains how eukaryotic messenger RNAs are destroyed or selectively destabilized after they have been made, processed, exported, translated, localized, or inspected by quality-control systems. The main focus is cytoplasmic mRNA decay, including deadenylation, decapping, Xrn1-mediated 5′ decay, nonsense-mediated decay, ribosome-associated quality-control decay, regulated endonucleolytic cleavage, and the connection between mRNA decay, translation, stress, localization, infection, development, and disease. [Chapter 26](chapter1025.md) covers cap chemistry and cap-binding proteins; [Chapter 29](chapter1028.md) covers 3′ end formation and poly(A)-tail addition; [Chapter 31](chapter1030.md) covers nuclear RNA surveillance; [Chapter 32](chapter1031.md) covers the enzymology of RNA decay machines; [Chapter 36](chapter1034.md) covers codon optimality and sequence-directed stability; [Chapter 38](chapter1036.md) covers stability programs in stress, development, immunity, aging, and disease.

## Executive Summary

Eukaryotic mRNA decay is the controlled loss of messenger RNA molecules through enzymatic removal of protective features, cleavage of the RNA body, or exonucleolytic digestion from a newly exposed end. The word "decay" can sound like passive chemical breakdown, but most cellular mRNA decay is an actively regulated pathway. A mature eukaryotic mRNA is protected by a 5′ cap, a 3′ poly(A) tail bound by poly(A)-binding proteins, translating ribosomes, and many RNA-binding proteins. Decay begins when those protective features are shortened, removed, rearranged, or bypassed.

The canonical cytoplasmic route is deadenylation-dependent decapping followed by 5′ to 3′ exonucleolytic degradation. In this model, deadenylase complexes shorten the poly(A) tail; loss or remodeling of poly(A)-binding protein changes the messenger ribonucleoprotein particle, or mRNP; decapping activators recruit and stimulate DCP2-containing decapping complexes; and Xrn1 digests the uncapped RNA body from the newly exposed 5′ end. This pathway is well supported by classic yeast experiments showing that deadenylation can precede decapping and 5′ decay of unstable mRNAs, and by later biochemical, genetic, and structural work on decapping factors (Muhlrad et al. 1994; Caponigro and Parker 1995; Mugridge et al. 2018; He and Jacobson 2023).

The canonical pathway is a teaching model, not a universal rule. Some mRNAs can be decapped before detectable deadenylation, some mRNAs are cut internally by endonucleases, and some quality-control pathways are triggered by translation termination, ribosome stalling, or the absence of a normal stop codon rather than by gradual tail shortening. Recent kinetic modeling and transcriptome-scale work emphasize that deadenylation rate and total mRNA decay rate are related but not interchangeable; measuring a short tail does not by itself prove that decay has already occurred (Czarnocka-Cieciura et al. 2024). Likewise, evidence that decapping-triggered degradation can be partly independent of initial deadenylation warns against treating the deadenylation-to-decapping sequence as obligatory for every transcript in every cell state (Audebert et al. 2024).

Nonsense-mediated decay, or NMD, is a translation-linked surveillance and regulatory pathway that preferentially destabilizes mRNAs in which translation termination occurs in an abnormal context. A typical textbook example is an mRNA with a premature termination codon generated by mutation or alternative splicing. When a ribosome terminates too early, surveillance factors including UPF proteins can distinguish the event from ordinary termination and promote mRNA decay. In mammals, exon junction complexes downstream of a stop codon can strengthen NMD, but NMD cannot be reduced to a single "downstream exon junction complex rule." Long 3′ untranslated regions, inefficient termination, upstream open reading frames, alternative splicing, viral RNA features, and circular-RNA-related contexts can all intersect with NMD-like logic in specific systems (Kishor et al. 2019; Garcia-Moreno and Romao 2020; Boo et al. 2024).

No-go decay and nonstop decay are ribosome-associated mRNA quality-control pathways. No-go decay responds to ribosomes that stall during elongation, for example at strong RNA structures, damaged RNA, problematic nascent peptide sequences, rare codon clusters, or collision-prone regions. Nonstop decay responds to translation complexes that reach the 3′ end of an mRNA without encountering a normal stop codon. These pathways protect proteostasis as well as mRNA integrity, because a defective mRNA can produce incomplete or harmful polypeptides. The same ribosome state can recruit protein quality-control factors, mRNA cleavage or decay factors, and ribosome rescue systems; therefore mRNA decay and protein quality control should be discussed together rather than as independent cleanup pathways (Alagar Boopathy et al. 2023; Gao et al. 2023).

Regulated endonucleolytic decay provides a second major entry point. An endonuclease cleaves inside the mRNA, creating fragments that are finished by exonucleases. Endonucleolytic decay can be triggered by microRNAs, small interfering RNAs, innate immune nucleases, stress responses, endoplasmic reticulum-associated quality control, or transcript-specific regulatory proteins, depending on organism and context. Endonucleases can bypass the need to remove the original cap or tail first, but endonucleolytic cleavage still usually depends on substrate recognition through RNP state, translation, localization, or signaling.

Eukaryotic mRNA decay is inseparable from translation, localization, and disease. Translation protects many mRNAs by engaging the cap and poly(A) tail in productive initiation cycles, but translation can also expose problems such as premature termination, ribosome stalling, frameshifts, or nonstop translation. Stress granules and processing bodies, or P-bodies, contain mRNAs and decay or repression factors, but visible granule localization does not prove that an mRNA is being degraded at that moment. Disease links arise when decay factors, decapping enzymes, surveillance pathways, or translation-quality-control systems are mutated or chronically misregulated. For example, biallelic NUDT2 variants defective in mRNA decapping have been linked to neurodevelopmental disease, and NMD can either suppress or expose cancer-relevant effects depending on whether it removes harmful mutant mRNAs, limits tumor-promoting transcripts, or masks neoantigen-producing transcripts (Husain et al. 2024; Nogueira et al. 2021).

## Concept Inventory

- **mRNA:** a messenger RNA that can be translated into protein. In eukaryotes, a typical mature mRNA has a 5′ cap, a coding sequence, untranslated regions, and a 3′ poly(A) tail. These features are not merely labels on a diagram; they are binding platforms for proteins that control translation, localization, and stability.
- **Poly(A) tail:** a stretch of adenosines added to the 3′ end of many eukaryotic mRNAs. Poly(A)-binding protein, often abbreviated PABP, coats the tail and helps promote translation and stability. Deadenylation is enzymatic shortening of the poly(A) tail. Deadenylation can be the first irreversible-looking step toward decay, but it can also be reversible or regulatory when cytoplasmic polyadenylation, tail protection, or storage pathways intervene.
- **Decapping:** removal of the 5′ cap or conversion of the cap into an RNA end compatible with decay. The major cytoplasmic decapping enzyme in many eukaryotes is a DCP2-containing complex, with DCP1 and multiple activators controlling activity and substrate choice. DCP2 is not simply a free enzyme that attacks every cap; decapping normally requires RNP remodeling and activation by cofactors.
- **Xrn1:** a conserved cytoplasmic 5′ to 3′ exonuclease that degrades RNAs with accessible 5′ ends. Xrn1 normally cannot begin on an intact capped mRNA. Xrn1 becomes effective after decapping or after endonucleolytic cleavage creates a suitable 5′ monophosphate.
- **Nonsense-mediated decay:** a translation-dependent mRNA surveillance and regulatory pathway that recognizes abnormal termination contexts. A premature termination codon is a stop codon that occurs upstream of the normal stop codon and can produce a truncated protein if the mRNA is translated. NMD often reduces abundance of premature-termination-codon-containing mRNAs, but NMD also regulates some apparently normal transcripts.
- **No-go decay:** mRNA decay triggered by ribosome stalling during elongation. Nonstop decay, or NSD, is mRNA decay triggered when a ribosome translates through the coding region and reaches the poly(A) tail or RNA end without a normal stop codon. Ribosome-associated quality control is a broader term for mechanisms that rescue stalled ribosomes and dispose of defective nascent polypeptides and mRNAs.
- **Processing body:** a cytoplasmic mRNP granule enriched for translationally repressed mRNAs and decay-related factors. A stress granule is a cytoplasmic granule enriched for translation-initiation-stalled mRNAs and RNA-binding proteins during stress. Neither term should be equated automatically with active degradation; granules can represent storage, triage, repression, remodeling, or decay-linked states.

## What to Know Before Reading This Chapter

RNA decay depends on RNA end chemistry. A eukaryotic mRNA cap protects the 5′ end from ordinary 5′ exonucleases, and the poly(A) tail helps recruit proteins that protect the 3′ end and support translation. [Chapter 2](chapter1002.md) explains 5′ and 3′ polarity; [Chapter 26](chapter1025.md) explains cap chemistry; [Chapter 29](chapter1028.md) explains polyadenylation.

Translation termination is an information-rich event. When a ribosome reaches a stop codon, release factors, nearby RNA-binding proteins, exon junction complexes, poly(A)-binding protein, RNA structure, and the length and composition of the 3′ untranslated region can influence whether termination is interpreted as normal. NMD uses this interpretive context rather than reading the word "nonsense" as a property of the codon alone.

An mRNA in a cell is an mRNP. Proteins bind the cap, tail, coding sequence, untranslated regions, splice-junction marks, localization elements, and regulatory motifs. Decay factors therefore act on mRNP states. A sequence motif that destabilizes an mRNA in one cell type can be neutral or stabilizing in another if different RNA-binding proteins, translation rates, stress signals, or compartments are present.

Steady-state RNA abundance is not the same as decay rate. An mRNA can be abundant because it is transcribed rapidly, stable, or both. A low-abundance mRNA can be unstable, weakly transcribed, or cell-type restricted. Direct claims about mRNA decay require evidence such as metabolic labeling, transcriptional shutoff with controls, poly(A)-tail measurement, cap-state analysis, RNA-end mapping, ribosome profiling, reporter assays, or acute perturbation of decay factors.

## 35.1. Deadenylation and poly(A)-tail control

Deadenylation is shortening of the 3′ poly(A) tail. The main object in this subsection is the mature cytoplasmic mRNA tail, not the nuclear polyadenylation reaction that initially creates the tail. A newly processed mRNA often leaves the nucleus with a poly(A) tail bound by poly(A)-binding proteins. These proteins help stabilize the mRNP and support translation initiation by cooperating with cap-binding and initiation factors. When deadenylases shorten the tail, fewer PABP molecules can bind, and the mRNA can shift toward translational repression, decapping, storage, or 3′ decay.

![Figure 35.1. Cytoplasmic mRNA Fate Map](../assets/figures/chapter1033_figure1.png)

**Figure 35.1. Cytoplasmic mRNA Fate Map.** A mature eukaryotic mRNA is protected by a cap, poly(A)-binding proteins, translating ribosomes, and transcript-specific RNA-binding proteins. Decay begins when these protective features are removed, remodeled, or bypassed by surveillance pathways. Depending on the cellular context, the same mRNA can continue translation, enter storage, undergo deadenylation-dependent decapping, be cut internally by an endonuclease, trigger nonsense-mediated decay after abnormal termination, or activate ribosome-associated quality control after stalling or nonstop translation.

![Figure 35.2. Five Entry Points into mRNA Decay](../assets/figures/chapter1033_figure2.png)

**Figure 35.2. Five Entry Points into mRNA Decay.** Eukaryotic mRNA decay pathways differ in the feature that licenses decay: a shortened tail, activated decapping complex, abnormal termination context, stalled or nonstop ribosome, or targeted internal cleavage. Each panel shows the same generic mRNA architecture — cap, coding sequence, 3′ UTR, and poly(A) tail — with the pathway-specific trigger and the downstream exonucleolytic finishing route highlighted, so differences in pathway logic rather than in mRNA structure are immediately apparent.

**Table 35.1. Major mRNA Decay Pathways and Diagnostic Evidence.** Overview of the principal cytoplasmic mRNA decay routes, their defining triggers, core molecular factors, the initial RNA event that commits the transcript to decay, the exonucleolytic route that finishes degradation, the evidence standard required for a strong mechanistic claim, and the most common interpretive error seen in the literature.

| Pathway | Trigger or substrate feature | Core factors | Initial RNA event | Finishing nuclease route | Strong evidence needed | Common overinterpretation |
| --- | --- | --- | --- | --- | --- | --- |
| **Deadenylation-dependent decay** | Poly(A)-tail shortening; reduced PABP occupancy | CCR4-NOT, PAN2-PAN3, DCP1-DCP2, Xrn1 | Tail shortening followed by decapping | Xrn1 (5′→3′); exosome (3′→5′) | Decay kinetics paired with tail-length assay; deadenylase mutant rescue | Short tail equals completed decay |
| **Decapping-first or decapping-triggered decay** | Translational repression; RNP remodeling without prior full deadenylation | DCP1-DCP2, EDC factors, Pat1, Lsm1-7, DDX6 | Cap hydrolysis | Xrn1 (5′→3′) | Cap-state assay; DCP2 catalytic mutant; Xrn1-dependent decay-fragment mapping | Deadenylation always precedes decapping |
| **NMD** | Abnormal termination context; premature stop codon; downstream exon junction complex | UPF1, UPF2, UPF3, SMG factors | UPF1-dependent mRNP remodeling and decay activation | Exonuclease and endonuclease routes | PTC reporter; UPF1 ATPase-mutant rescue; ribosome profiling | All PTC-containing mRNAs are eliminated; NMD requires downstream EJC |
| **NGD** | Ribosome stalling during elongation; RNA structure; rare codons; collision | Collision sensors, ribosome-quality-control factors, endonuclease | Endonucleolytic cleavage near stall site | Xrn1 (5′→3′); exosome (3′→5′) | Stall reporter; RNA-end mapping; collision-factor genetic requirement | Ribosome pausing alone equals no-go decay |
| **NSD** | Missing stop codon; poly(A)-tail translation | Ribosome-rescue factors, exosome, Ski complex | Exosome-mediated 3′-to-5′ digestion; ribosome rescue | Exosome (3′→5′); Xrn1 after ribosome splitting | Nonstop reporter; ribosome-rescue factor dependence; poly(A)-tail translation evidence | NSD and NMD use identical machinery |
| **Small RNA-guided slicing** | Guide-complementary site; slicing-competent Argonaute | Slicing-competent Argonaute; siRNA or plant miRNA guide | Internal endonucleolytic cleavage | Xrn1 (5′→3′); exosome (3′→5′) for generated fragments | Slicing-competent Argonaute; cleavage-site mapping; catalytic-mutant rescue | All miRNA effects involve slicing |
| **ER/stress-linked endonucleolytic decay** | ER stress; unfolded protein response; RNase L activation | IRE1, RNase L, ER-associated decay factors | Internal endonucleolytic cleavage | Exonuclease cleanup of newly generated ends | Nuclease dependence; RNA-end mapping; stress-specific fragment accumulation | ER stress degrades all cytoplasmic mRNAs equally |

**Table 35.2. Selected Factors in Eukaryotic mRNA Decay.** Key proteins and complexes that execute or regulate mRNA decay, with their primary roles, the logic that directs them to substrates, the experimental approaches that have established their function, and representative supporting work.

| Factor or complex | Primary role | Examples of recruitment logic | Assays that support function | Citation anchors |
| --- | --- | --- | --- | --- |
| **PAN2-PAN3** | Initial poly(A)-tail trimming | Recruited to PABP-bound mRNPs on long tails; often acts before CCR4-NOT | Poly(A)-tail sequencing in PAN2 mutants; deadenylation kinetics | Liu et al. 2023 |
| **CCR4-NOT** | Central deadenylase and regulatory scaffold | Recruited by RNA-binding proteins, microRNA effectors, and codon-optimality pathways | Deadenylation kinetics; NOT subunit depletion; structural studies | Mugridge et al. 2018; Liu et al. 2023 |
| **PABP/PABPC** | Poly(A)-tail binding; promotes translation and mRNP stability | Coats poly(A) tail; cooperates with cap-binding and initiation factors | Poly(A)-tail assay; PABP mutant decay analysis; ribosome profiling | Caponigro and Parker 1995 |
| **LARP1/LARP4** | Poly(A)-tail protection and translation modulation | LARP1 engages 5′ TOP mRNAs under nutrient signals; LARP4 stabilizes subsets of mRNAs | mRNA stability assays; LARP depletion; PABP co-immunoprecipitation | Mattijssen et al. 2021 |
| **DCP1-DCP2** | Catalytic decapping of the mRNA 5′ cap | Activated by cofactors in repressed or deadenylated mRNPs; requires RNP remodeling | In vitro decapping assay; DCP2 catalytic mutant; cap-state analysis | He and Jacobson 2023 |
| **Pat1/Lsm1-7/DDX6** | Decapping activation; decapping complex assembly | Pat1 bridges deadenylation and decapping complexes; Lsm1-7 binds oligoadenylated tails | Pat1 depletion; P-body imaging; co-immunoprecipitation | Ozgur et al. 2010 |
| **Xrn1** | 5′-to-3′ exonucleolytic degradation of decapped RNA | Engages 5′ monophosphate after decapping or endonucleolytic cleavage; blocked by intact cap | Xrn1 deletion phenotype; decay-fragment accumulation; in vitro exonuclease assay | Mugridge et al. 2018 |
| **UPF1/UPF2/UPF3 and SMG factors** | NMD surveillance; recognition of abnormal termination | Recruited at premature stop codons or downstream EJCs; UPF1 ATPase-dependent remodeling | NMD reporter; UPF1 ATPase-mutant rescue; SMG factor depletion | Kishor et al. 2019 |
| **Argonaute** | Small RNA-guided target repression or slicing | Guide RNA directs Argonaute to complementary mRNA; slicing requires high complementarity | Argonaute immunoprecipitation; slicing assay; catalytic-mutant rescue | Gao et al. 2023 |
| **Ribosome-associated quality-control factors** | Ribosome rescue; mRNA decay after stalling or nonstop translation | Recruited to stalled or collided ribosomes; couple mRNA cleavage to nascent-chain ubiquitylation | Collision profiling; ribosome-rescue mutants; RNA-end mapping | Alagar Boopathy et al. 2023 |
| **NUDT2** | Cap hydrolysis; Nudix decapping enzyme | Acts on subsets of capped or cap-like RNA structures; clinically linked to neurodevelopmental disease | Variant functional assay; in vitro hydrolysis; RNA profiling in patient cells | Husain et al. 2024 |

**Table 35.3. Disease and Stress Interpretation Matrix.** Summary of contexts in which mRNA decay pathways have disease or stress relevance, comparing potential protective and harmful effects and the evidence standard needed to establish causality.

| Context | Decay pathway involved | Possible protective effect | Possible harmful effect | Evidence needed for causality |
| --- | --- | --- | --- | --- |
| **Inherited premature termination codon** | NMD | Prevents production of toxic truncated protein | Destroys mRNA that could yield a partially functional protein, worsening loss-of-function disease | PTC reporter; UPF factor dependence; truncated protein analysis; phenotype rescue |
| **Cancer NMD alteration** | NMD | Removes nonsense-mutant transcripts from tumor-suppressor genes | Masks neoantigen-producing mRNAs; eliminates regulatory or stress-protective targets | NMD factor expression levels; isoform profiling; neoantigen immune assay |
| **Neurodevelopmental decapping-factor variant** | Decapping; cap hydrolysis | None established in this context | Impaired cap removal causes developmental mRNA dysregulation | Variant functional assay; RNA profiling in patient cells; phenotype rescue |
| **Viral RNA infection** | NMD/UPF1-mediated decay; viral subversion of host decay | Host NMD can degrade viral mRNAs and limit replication | Viral proteins can subvert NMD or other decay factors to stabilize viral RNAs | Viral RNA stability in UPF1-depleted cells; viral protein interaction assays |
| **ER stress** | ER-associated endonucleolytic decay | Clears misfolded-protein mRNAs to relieve ER burden | Excessive mRNA decay can impair stress-recovery translation | IRE1 nuclease dependence; RNA-end mapping; stress time course with factor rescue |
| **Neuronal mRNA localization** | Localization-coupled stability | mRNA stored stably for on-demand local translation at synapses | Premature decay prevents local protein synthesis and disrupts neuronal function | Live imaging; local translation assay; decay-factor perturbation in neurites |

The simplest causal model has four steps. First, an mRNA carries a poly(A) tail that supports translation and protects the 3′ end. Second, regulatory proteins, microRNA-associated factors, codon optimality pathways, developmental cues, or stress signals recruit deadenylase complexes. Third, tail shortening reduces PABP occupancy and changes the mRNP's physical and regulatory state. Fourth, the mRNA becomes more likely to be decapped and degraded, although some transcripts enter storage or regulated re-polyadenylation instead. Reviews of 5′ to 3′ decay and poly(A)-tail regulatory proteins support this basic framework (Mugridge et al. 2018; Mattijssen et al. 2021; Liu et al. 2023).

![Figure 35.3. Translation State Controls Stability and Granule Partitioning](../assets/figures/chapter1033_figure3.png)

**Figure 35.3. Translation State Controls Stability and Granule Partitioning.** Translation can protect a normal mRNA, expose an abnormal termination event, reveal ribosome stalling, or prevent stress-induced granule localization. Stress granule or P-body localization is an mRNP state, not automatic proof of degradation; a stress-granule-localized mRNA may be stored, triaged, returned to translation, or eventually degraded depending on the cellular context and the duration and type of stress.

Two major deadenylase systems are commonly emphasized in eukaryotic mRNA decay: the PAN2-PAN3 complex and the CCR4-NOT complex. PAN2-PAN3 often participates in early trimming of long tails, whereas CCR4-NOT is a central deadenylase and regulatory platform recruited by many RNA-binding proteins and small-RNA pathways. The CCR4-NOT complex is not just a nuclease; it is a multi-subunit scaffold that can connect deadenylation to translational repression, decapping activation, protein recruitment, and signaling. Exact subunit names and dominance vary by organism and cell type, so a claim about "the deadenylase" should specify the complex and system.

A concrete example is a short-lived yeast mRNA such as MFA2, in which classic experiments showed that deadenylation of an unstable transcript could be followed by decapping and 5′ to 3′ digestion (Muhlrad et al. 1994). That work helped define deadenylation-dependent decapping as a core pathway. A second example is poly(A)-binding protein itself. PABP can influence both deadenylation and decapping, and yeast genetics showed that PABP has multiple roles in the timing and coupling of these steps (Caponigro and Parker 1995). These examples are useful because they make the tail more than a length measurement: the tail is a protein-coated regulatory platform.

Tail-control proteins can also stabilize or specialize mRNAs. LARP1 and LARP4 are La-related RNA-binding proteins that interact with PABP and influence poly(A)-tail protection, translation, and stability in transcript-specific ways (Mattijssen et al. 2021). LARP1 is often discussed in relation to terminal oligopyrimidine motif-containing mRNAs and nutrient-responsive translation control, whereas LARP4 family proteins are often framed as stabilizing factors for subsets of mRNAs. These proteins show why the phrase "short tail equals decay" is too crude. Tail length, PABP occupancy, 5′ cap context, translation initiation state, and bound regulatory proteins jointly determine fate.

The evidence basis for deadenylation includes Northern blotting of poly(A)-tail length, RNase H and oligo(dT)-directed tail assays, poly(A)-tail sequencing, metabolic labeling, reporter genes, deadenylase mutants, biochemical assays, and mathematical modeling of decay kinetics. Each method has limits. Bulk RNA-seq usually does not report tail length. Tail-length assays can be biased by library construction, transcript isoforms, and 3′ end heterogeneity. Inhibiting transcription to measure decay can induce stress. Depleting a deadenylase can change transcription, translation, and cell state indirectly.

Recent modeling highlights a central caveat: deadenylation rate and decay rate are coupled but not identical. Some transcripts deadenylate rapidly and decay rapidly; others deadenylate without immediate destruction, or decay by routes that do not begin with detectable tail shortening. Kinetic models that infer deadenylation rates from time-resolved data reveal a complex relationship between tail shortening and decay outcomes (Czarnocka-Cieciura et al. 2024). This complexity should not be treated as a contradiction of the canonical model. It means the canonical model is one branch in a larger decision network.

Do not overgeneralize across compartments and RNA classes. Nuclear RNAs can be oligoadenylated as decay signals, plant and organellar tails can have different meanings, and cytoplasmic mRNAs can undergo regulated cytoplasmic polyadenylation during development or neuronal activation. [Chapter 29](chapter1028.md) covers tail addition and alternative polyadenylation; [Chapter 36](chapter1034.md) covers sequence and codon features that influence deadenylation and stability.

## 35.2. Decapping and 5-prime decay

Decapping is the step that removes the protective 5′ cap from an mRNA or otherwise converts the cap-bearing end into a decay-competent end. For most cytoplasmic eukaryotic mRNAs, the cap is a 7-methylguanosine-linked structure that promotes translation initiation and blocks direct access by Xrn1. After decapping, Xrn1 can engage the exposed 5′ monophosphate and degrade the RNA body from 5′ to 3′. [Chapter 26](chapter1025.md) treats cap chemistry and decapping enzymes in broader detail; this section focuses on decapping as a commitment point in mRNA decay.

The DCP2-containing decapping complex is the central cytoplasmic decapping machine in many eukaryotes. DCP2 supplies catalytic activity, DCP1 stimulates and organizes the complex, and additional activators such as EDC proteins, Pat1, DDX6-family helicases, and Lsm1-7 factors help recruit, remodel, or activate decapping on specific mRNP states. Reviews of eukaryotic decapping emphasize that decapping factors control both molecular catalysis and substrate selection (Mugridge et al. 2018; Vidya and Duchaine 2022; He and Jacobson 2023).

A causal decapping sequence can be described as follows. First, an mRNA becomes less engaged in productive translation, often after deadenylation or regulatory repression. Second, decapping activators bind the mRNP, sometimes through interactions with deadenylation factors, PABP loss, microRNA effector complexes, or RNA-binding proteins. Third, DCP2 is activated in a conformation and RNP context that allows cap hydrolysis. Fourth, Xrn1 degrades the decapped RNA. Pat1b provides an example of coupling: human Pat1b can connect deadenylation with decapping and influence P-body assembly, illustrating how a cofactor can link enzymatic steps and cytoplasmic RNP organization (Ozgur et al. 2010).

The pathway is often shown as a conveyor belt from deadenylation to decapping to Xrn1 digestion. That diagram is useful, but current evidence requires a more flexible view. A 2024 study reported that RNA degradation triggered by decapping can be largely independent of initial deadenylation in tested contexts (Audebert et al. 2024). This does not erase deadenylation-dependent decay; instead, it shows that mRNAs can enter 5′ decay through multiple gates. Some transcripts may be decapped after tail shortening, others may be decapped after translational repression or endonucleolytic events, and still others may be primarily degraded from the 3′ side or by quality-control pathways.

Decapping is also broader than DCP2. Nudix hydrolases and other cap-removing enzymes can act on subsets of RNAs or cap-like structures. NUDT2 is clinically relevant because biallelic NUDT2 variants defective in mRNA decapping have been reported to cause a neurodevelopmental disease (Husain et al. 2024). The disease example should be interpreted carefully. It supports the importance of cap-removal biology in human development, but it does not mean that all neurodevelopmental phenotypes caused by RNA-decay genes arise from the same transcript set or pathway.

Plants and other eukaryotes add additional complexity. RNA uridylation can promote decay in plants, and uridylated tails can influence decapping, exonuclease access, or 3′ decay depending on RNA class and organism (de Almeida et al. 2018). Helicases can also influence decay by remodeling structured mRNAs or RNPs before decapping or exonuclease engagement (Khemici and Linder 2018). These examples reinforce a general rule: decapping is not only a chemical event at the cap; it is an RNP remodeling decision.

Experimental evidence for decapping and 5′ decay includes cap-state assays, accumulation of decay intermediates in Xrn1-deficient cells, decapping-enzyme mutants, in vitro decapping assays, structural studies of decapping complexes, reporter mRNAs with defined caps and tails, and transcriptome-wide measurements of decay fragments. Assay interpretation has several traps. A rise in an RNA after DCP2 depletion may reflect direct decapping dependence, indirect stress, altered transcription, or changed translation. P-body accumulation of a transcript does not prove decapping. A 5′ fragment can arise from endonucleolytic cleavage rather than decapping unless the end chemistry and cleavage position are mapped.

Boundary cases are common. Some mRNAs can be stabilized but translationally repressed, some decapped RNAs may persist long enough to be detected, and some endonuclease-generated fragments feed Xrn1 without prior removal of the original cap. Viral RNAs may carry unusual caps, protein-linked ends, internal ribosome entry sites, or structured untranslated regions that alter access by host decay factors. [Chapter 115](chapter1109.md) and [Chapter 117](chapter1111.md) treat viral RNA strategies in detail; here the main principle is that cap-dependent decay logic must be tested against each RNA end architecture.

## 35.3. Nonsense-mediated decay

Nonsense-mediated decay, or NMD, is a translation-dependent pathway that destabilizes mRNAs when translation termination occurs in an abnormal context. The name comes from "nonsense" mutations, which introduce premature stop codons, but modern NMD is not limited to genetic nonsense mutations. NMD also regulates many natural transcripts, including alternatively spliced isoforms, transcripts with upstream open reading frames, long 3′ untranslated regions, retained introns, or other features that alter termination context.

The basic problem NMD solves is discrimination between normal and abnormal termination. A ribosome terminating at the correct stop codon usually lies near factors that promote efficient termination and recycling, including poly(A)-binding protein in many mRNP configurations. A ribosome terminating far upstream of the normal 3′ end, upstream of a downstream exon junction complex, or in an otherwise inefficient termination environment can recruit surveillance factors. UPF1 is a central RNA helicase-like ATPase in NMD; UPF2 and UPF3-family proteins help connect UPF1 to downstream marks and activation pathways in many systems. Additional factors regulate UPF1 phosphorylation, remodeling, and handoff to decay machinery.

A concrete mammalian example is an alternatively spliced mRNA that includes a poison exon. A poison exon is an exon whose inclusion introduces a premature termination codon. Translation of the poison-exon isoform causes termination upstream of the normal stop codon and often upstream of exon junction complexes deposited by splicing. NMD then reduces abundance of that isoform, creating a coupling between alternative splicing and gene-expression control. Reviews of alternative splicing coupled to NMD emphasize that this logic is used in autoregulation of splicing factors and developmental gene-expression programs, not only in elimination of mutant mRNAs (Garcia-Moreno and Romao 2020).

![Figure 35.4. Nonsense-Mediated Decay as an Abnormal-Termination Decision](../assets/figures/chapter1033_figure4.png)

**Figure 35.4. Nonsense-Mediated Decay as an Abnormal-Termination Decision.** Normal termination near a poly(A)-binding-protein-rich 3′ end should be contrasted with four context-dependent NMD-prone architectures: a downstream exon junction complex, a long 3′ untranslated region, an upstream open reading frame, and a retained intron. The abnormal-termination branch recruits UPF recognition and SMG-dependent remodeling, then divides between SMG6 endonucleolysis and deadenylation/decapping before exonucleolytic clearance. A visible escape branch should prevent the common false rule that every transcript with one of these features is obligatorily destroyed.

The exon junction complex rule is useful but incomplete. In many mammalian transcripts, a stop codon more than a modest distance upstream of a downstream exon junction complex is more likely to trigger NMD. However, NMD also depends on translation efficiency, 3′ untranslated region length and composition, proximity of PABP, RNA-binding proteins, stop-codon context, cell type, and the kinetics of termination. Yeast NMD lacks the same exon junction complex logic because budding yeast splicing and mRNP architecture differ from those of mammals. Therefore a reader should not define NMD as "degradation of mRNAs with downstream exon junction complexes"; that is one mechanistic route in a broader surveillance family (Kishor et al. 2019).

NMD can act as quality control and regulation at the same time. In inherited disease, NMD can reduce production of truncated proteins from premature-termination-codon-containing alleles, which may be protective when truncated proteins are toxic. Conversely, NMD can worsen loss-of-function disease by destroying an mRNA that could produce a partially functional protein. In cancer, NMD can remove mRNAs from tumor-suppressor genes with nonsense mutations, shape immune recognition by controlling neoantigen-producing transcripts, and regulate nonmutant transcripts involved in stress and growth. This duality explains why reviews describe NMD as having context-dependent, sometimes bipolar roles in cancer (Nogueira et al. 2021).

NMD also intersects with infection. Some viral RNAs are targeted by UPF1-mediated decay pathways, whereas viruses can evolve structures, proteins, translation strategies, or replication compartments that evade or manipulate NMD. Coronavirus infection provides an example of interplay between a cytoplasmic RNA virus and NMD, with viral RNA biology and host decay machinery influencing each other (Wada et al. 2018; May and Simon 2021). A viral RNA should not be assumed to follow host mRNA rules simply because it is translated; viral replication complexes, RNA structures, caps, poly(A) tails, and protein interactions can change surveillance exposure.

Recent work extends NMD-related thinking to circular RNAs in particular contexts. Circular RNAs lack ordinary 5′ and 3′ ends, so their decay cannot proceed by the same end-first logic as linear mRNA decay. A 2024 study reported that circular RNAs can trigger nonsense-mediated mRNA decay (Boo et al. 2024). The result is a useful reminder that NMD is not merely a linear RNA disposal pathway; it is a translation-context surveillance system whose targets and consequences can include unusual RNA architectures. Final-reference expansion should add more circular RNA-specific background references from [Chapter 94](chapter1089.md) to contextualize how circular RNA translation and surveillance are detected.

Experimental evidence for NMD includes premature-termination-codon reporters, splice isoform measurements, UPF factor depletion or mutation, rescue with wild-type and ATPase-defective factors, ribosome profiling, RNA-seq after acute inhibition, phosphoprotein analysis, and decay-fragment mapping. Interpretation requires care. UPF1 participates in multiple RNA processes, so UPF1 depletion can produce indirect changes. Translation inhibitors can suppress NMD but also globally alter cell physiology. Reporter mRNAs can overrepresent strong NMD rules that are weaker in endogenous mRNPs. Strong NMD claims connect termination context, translation, NMD factor dependence, and decay rather than relying on steady-state abundance alone.

## 35.4. No-go, nonstop, and ribosome-associated decay pathways

No-go decay and nonstop decay are mRNA quality-control pathways built around ribosome state. No-go decay, or NGD, responds when ribosomes stall during elongation. Nonstop decay, or NSD, responds when translation reaches the end of an mRNA without encountering a normal stop codon. These pathways are ribosome-associated because the cell detects a translation problem, not merely an RNA sequence defect visible to a free nuclease.

The background bridge is translation elongation and termination. A healthy ribosome moves codon by codon along an mRNA, adds amino acids to a growing polypeptide, and terminates at a stop codon with release factors. Problems arise when the ribosome cannot move, collides with another ribosome, translates through a missing stop codon, reaches a truncated RNA end, or encounters a region that produces a defective nascent chain. The cell must rescue the ribosome, dispose of the incomplete protein, and often destroy the mRNA that caused the problem. [Chapter 71](chapter1066.md) treats ribosome quality control and proteostasis in more detail.

In NGD, ribosome stalling can be caused by strong RNA structures, damaged bases, rare-codon clusters, problematic nascent peptides, defective translation factors, or collisions created by high ribosome load. Stalled or collided ribosomes recruit quality-control proteins that split ribosomal subunits, ubiquitylate ribosomal proteins or nascent chains, and stimulate endonucleolytic cleavage or exonucleolytic decay of the mRNA. The mRNA fragment downstream or upstream of the stalled ribosome can then be degraded by Xrn1, the exosome, or other nucleases depending on the generated ends and organism. Reviews of no-go decay and ribosome quality control emphasize that mRNA decay and protein quality control are coordinated responses to one translation failure (Alagar Boopathy et al. 2023).

In NSD, the key abnormal feature is absence of a functional stop codon. A ribosome may translate into the poly(A) tail, producing a lysine-rich peptide segment, or reach the physical end of a truncated mRNA. The cell must rescue a ribosome that lacks the normal termination signal and prevent repeated production of aberrant proteins. NSD is conceptually distinct from NMD: NMD responds to termination at a stop codon in a suspicious context, whereas NSD responds to a missing stop codon or missing normal termination event. In real cells, boundaries can blur because truncated RNAs, frameshifts, poly(A)-tail translation, and ribosome collisions can activate overlapping quality-control factors.

![Figure 35.5. No-Go and Nonstop Decay: Ribosome Rescue, RNA Clearance, and Nascent-Chain Disposal](../assets/figures/chapter1033_figure5.png)

**Figure 35.5. No-Go and Nonstop Decay: Ribosome Rescue, RNA Clearance, and Nascent-Chain Disposal.** Three parallel state paths should distinguish an elongation stall that produces a collided-ribosome no-go substrate, a ribosome that reaches a truncated RNA end, and translation through a missing stop codon into a poly(A) tail. The no-go path can initiate local RNA cleavage, whereas the two nonstop paths reach an abnormal RNA end without requiring the same cleavage event. All three routes must coordinate ribosome rescue and subunit splitting with RNA-fragment clearance and nascent-chain disposal; factor-level collision sensing and proteostasis depth remain a handoff to [Chapter 71](chapter1066.md).

Argonaute-associated pathways add another layer. Small RNA-guided Argonaute proteins can repress translation, recruit deadenylation and decapping factors, or slice target RNAs when catalytic slicing is available. Recent reviews discuss Argonaute-dependent ribosome-associated protein quality control, showing that small-RNA pathways can connect target recognition, ribosome state, and nascent-protein quality control (Gao et al. 2023). This does not mean every microRNA target undergoes NGD; most metazoan microRNA effects are better understood as translational repression and deadenylation-promoted destabilization. The ribosome-associated quality-control label should be reserved for cases where ribosome stalling or collision is mechanistically demonstrated.

Endoplasmic reticulum-associated decay is an important context because many secretory and membrane-protein mRNAs are translated on ER-bound ribosomes. Stress at the ER, defective targeting, or stalled translocation can couple protein folding and mRNA decay pathways. A recent review of RNA decay pathways at the endoplasmic reticulum frames the ER as a spatial platform where mRNA decay, translation, protein targeting, and stress signaling intersect (Ottens et al. 2024). The ER example helps readers avoid thinking of cytoplasmic decay as uniformly diffuse.

Evidence for NGD and NSD often combines ribosome profiling, disome or collision profiling, reporter constructs with stall sequences or missing stop codons, RNA-end mapping, proteomics of nascent-chain quality control, genetic perturbation of rescue factors, and imaging or fractionation of ribosome-associated mRNAs. Each assay has limitations. Ribosome profiling can show protected fragments at a stall but cannot alone prove mRNA cleavage. Reporter stalls can be artificial. Collision signatures can arise secondarily from translation initiation changes. A strong claim links ribosome state to decay-factor recruitment, RNA cleavage or destabilization, and rescue of the phenotype by pathway-specific factors.

Comparative biology is helpful but must be used carefully. Bacteria have ribosome rescue systems such as tmRNA and alternative rescue factors, and some supplied references discuss bacterial ribosome rescue (Muller et al. 2021). These bacterial systems are not the same as eukaryotic NGD and NSD, but they illustrate a universal problem: translating cells need mechanisms to rescue ribosomes and eliminate defective mRNAs. [Chapter 34](chapter1032.md) treats bacterial turnover, and [Chapter 71](chapter1066.md) treats ribosome rescue and proteostasis across systems.

## 35.5. Regulated endonucleolytic decay and surveillance coupling

Regulated endonucleolytic decay begins with an internal cut in an mRNA. This strategy differs from deadenylation-dependent decapping because it does not require gradual removal of the original tail or cap before the RNA body becomes vulnerable. An internal cleavage generates two fragments: a downstream fragment with a newly exposed 5′ end and an upstream fragment with a newly exposed 3′ end. These fragments can then be degraded by 5′ to 3′ and 3′ to 5′ exonucleases.

Endonucleolytic decay is regulated because internal cleavage must be targeted. A cell cannot allow nonspecific endonucleases to cut translating mRNAs randomly. Targeting can be supplied by small RNAs, RNA-binding proteins, ribosome stalls, innate immune sensors, endoplasmic reticulum stress pathways, nuclear surveillance factors, or transcript-specific structures. The same chemistry, phosphodiester cleavage inside RNA, can produce very different biological outcomes depending on who recruits the nuclease and where cleavage occurs.

Small RNA-guided cleavage is a clear example. In RNA interference pathways with slicing-competent Argonaute proteins, a small RNA guides Argonaute to a complementary target, and the target can be cleaved internally. The cleaved fragments are then cleared by exonucleases. Metazoan microRNAs usually act through repression and deadenylation rather than slicing, because animal microRNA-target pairing is often imperfect; however, small interfering RNAs and plant microRNAs commonly use highly complementary targeting that can produce slicing. This boundary case matters because "small RNA-mediated decay" can mean deadenylation-dominated decay or direct endonucleolytic cleavage depending on the pathway.

Ribosome-associated quality-control pathways can also use endonucleolytic cleavage. A stalled ribosome or ribosome collision can mark a site near which the mRNA is cleaved, creating fragments for exonucleolytic cleanup. This links [Section 35.4](chapter1033.md) to regulated endonucleolytic decay. The ribosome is not merely a passive obstacle; it can provide the surveillance context that licenses cleavage.

Nuclear and cytoplasmic surveillance are coupled. Nuclear mRNA decay networks control nascent transcripts, retained introns, misprocessed RNAs, and mRNPs that fail maturation or export; cytoplasmic pathways control translated mRNAs and translation-linked defects. A recent review of nuclear mRNA decay emphasizes that nuclear pathways regulate gene expression rather than simply remove mistakes (Rambout and Maquat 2024). For [Chapter 35](chapter1033.md), the important handoff is that nuclear history can mark a cytoplasmic mRNA. Splicing can deposit exon junction complexes that influence NMD; 3′ end formation can determine tail length and UTR architecture; export history can influence bound proteins.

UPF1-mediated decay pathways provide a surveillance-coupling example beyond the narrowest NMD definition. UPF1 can participate in multiple decay routes connected to termination, 3′ UTR context, viral RNAs, and transcript-specific features (Lavysh and Neu-Yilik 2020; May and Simon 2021). The term "UPF1-mediated decay" should therefore be used with more precision than "NMD" when the evidence shows UPF1 dependence but not the full canonical NMD mechanism.

Regulated endonucleolytic decay also intersects with immune signaling and stress, although this chapter does not cover innate immunity exhaustively. RNase L, IRE1-dependent decay in the unfolded protein response, and related stress-regulated cleavage pathways can reduce specific RNA populations while generating fragments that influence signaling. The ER-associated decay review by Ottens et al. (2024) is especially relevant for stress-linked endonucleolytic logic at the ER. Final-reference expansion should add direct IRE1/RIDD and RNase L primary or review references before final release if those pathways are expanded beyond this overview.

Evidence for endonucleolytic decay requires RNA-end resolution. If an endonuclease cuts internally, one expects cleavage-site-proximal fragments, changes in end chemistry, or accumulation of fragments when exonucleases are impaired. RNA-seq read coverage alone may suggest a cleavage region, but library preparation can create apparent ends. Strong evidence combines nuclease perturbation, catalytic mutants, mapping of newly generated ends, rescue, and ideally biochemical or structural support for substrate recognition.

## 35.6. Crosstalk with translation, stress, localization, and disease

Eukaryotic mRNA decay is coupled to translation because the same mRNA features that promote translation often protect against decay. The cap recruits initiation factors; the poly(A) tail recruits PABP; the coding sequence determines ribosome traffic; untranslated regions recruit regulatory proteins; and translation termination provides surveillance information. A transcript's stability is therefore partly an emergent property of how efficiently and where it is translated. [Chapter 36](chapter1034.md) develops codon optimality and sequence-directed stability in detail; this section focuses on pathway crosstalk.

A productive translating mRNP can be stabilized by cap-dependent initiation and PABP-supported architecture. When initiation is inhibited, ribosomes run off, PABP occupancy can change, and translationally repressed mRNAs can become available for decapping, storage, or granule assembly. Conversely, translation can expose defects. A premature stop codon triggers NMD only after translation reaches it. A stall triggers NGD only when ribosome movement fails. A missing stop codon triggers NSD only during attempted translation. Thus translation protects normal mRNAs and reveals abnormal ones.

Stress changes this balance. Heat shock, oxidative stress, nutrient limitation, viral infection, ER stress, and immune signaling can globally repress translation while selectively preserving or inducing translation of stress-response mRNAs. Stress granules contain translation-initiation-stalled mRNPs, while P-bodies are enriched for translationally repressed and decay-factor-associated mRNPs. The two granule types exchange components and mRNAs, but visible granules are not synonymous with decay. A stress-granule-localized mRNA may be stored, triaged, later returned to translation, or eventually degraded. Recent work showing that ribosome association can inhibit stress-induced localization of gene mRNAs to stress granules illustrates that translation state itself can determine whether an mRNA enters granule-like compartments (Helton et al. 2025).

Localization makes decay spatial. Neurons, oocytes, embryos, migrating cells, and polarized epithelial cells transport mRNAs to specific regions where local translation is needed. Localized mRNAs often remain translationally repressed during transport and then activate at the destination. Decay must be coordinated with this itinerary: premature decay would prevent local protein synthesis, whereas excessive stabilization could mislocalize protein production. Reviews of intracellular mRNA transport and localized translation emphasize that localization, translation, and stability are jointly regulated rather than sequentially independent (Das et al. 2021). [Chapter 74](chapter1069.md) treats localization signals and transport granules in more detail.

Metabolism also influences mRNA decay through translation. Cellular energy state, nutrient availability, amino acid supply, and signaling pathways change initiation, elongation, ribosome pausing, and stress responses. Reviews on metabolism-translation crosstalk emphasize that translation is both a consumer of metabolic resources and a sensor of cellular state (Biffo et al. 2024). Because many decay pathways are translation-linked, metabolic shifts can indirectly alter decay by changing ribosome traffic, initiation rates, tail protection, and stress-granule dynamics.

Disease links arise through several mechanisms. A mutation can create a premature termination codon and make an mRNA an NMD substrate. A mutation in a decay factor can impair turnover of many transcripts. A cancer cell can benefit from altered NMD or deadenylation programs that reshape stress tolerance, antigen presentation, or expression of growth regulators. A neurodevelopmental disorder can arise when decapping, ubiquitination-translation crosstalk, or ribosome-associated quality control fails during development (Husain et al. 2024; Elu et al. 2024). The disease-relevant question is not only whether RNA abundance changes, but which molecular defect links decay to cell physiology.

Noncoding RNA crosstalk also affects decay programs. MicroRNAs, long noncoding RNAs, circular RNAs, and RNA-binding proteins can compete for factors, scaffold decay machinery, alter translation, or change localization. Some claims in this area are strongly supported; others are correlative. Reviews of noncoding RNA crosstalk in brain health and disease provide useful context but should not be used to infer direct decay mechanisms without pathway-specific evidence (Mehta et al. 2021).

Assay limitations are especially important in disease and stress studies. Patient cells may differ in cell type composition, developmental state, compensation, and stress history. Knockdown of a decay factor can trigger secondary transcriptional and immune responses. Translation inhibitors used to probe NMD or stress granules can themselves alter decay. Disease-associated variants in broad decay factors can affect mRNA decay, noncoding RNA decay, translation, and protein quality control simultaneously. A robust disease mechanism links genotype, molecular pathway, transcript or protein targets, cellular phenotype, and rescue.

## Experimental Foundations and Evidence

The strongest mRNA decay studies separate synthesis, processing, translation, and degradation. Metabolic RNA labeling measures the appearance and disappearance of newly synthesized RNA. Transcriptional shutoff can estimate decay rates, but shutoff conditions often perturb physiology. Poly(A)-tail sequencing measures tail length and tail-length distributions but does not by itself prove decay. Cap analysis and 5′ end mapping distinguish decapping from endonucleolytic cleavage. Ribosome profiling and collision profiling reveal translation state but require paired RNA decay measurements to assign causality.

Reporter assays are powerful because they isolate a feature such as a premature stop codon, AU-rich element, stall sequence, poly(A)-tail length, or 3′ UTR. Their weakness is the same isolation. A reporter may lack native chromatin history, splicing context, localization, mRNP composition, RNA modifications, or transcript dosage. Endogenous validation is therefore essential for broad claims.

Genetic perturbation is informative when it distinguishes catalytic activity from scaffolding. A nuclease-dead DCP2 mutant, UPF1 ATPase mutant, deadenylase catalytic mutant, or rescue with a wild-type factor can clarify mechanism. Depletion alone is less specific because decay factors often have multiple partners and indirect effects.

Structural and biochemical evidence explains molecular possibility. Structures of decapping factors and decay complexes show how enzymes recognize caps, cofactors, and RNA, while in vitro assays test catalytic requirements. Cellular evidence is still needed because purified systems may lack the RNP context that makes a transcript a physiological substrate.

> **Box 35.1. Common Assay Traps in mRNA Decay**
>
> - Steady-state abundance is not decay rate. A low-abundance mRNA may be weakly transcribed, cell-type restricted, or rapidly decayed; distinguishing these requires kinetic or perturbation evidence.
> - Tail length is not decay by itself. A short poly(A) tail indicates deadenylation has occurred but does not prove that the mRNA has been destroyed or will be destroyed imminently.
> - P-body or stress-granule localization is not proof of degradation. Granule-resident mRNAs may be stored, triaged, or returned to translation; direct evidence of RNA loss or cleavage is required.
> - UPF1 dependence is not automatically canonical NMD. UPF1 participates in multiple RNA processes; demonstrating the full NMD mechanism requires connecting termination context, translation, UPF factor dependence, and decay.
> - Ribosome pausing is not automatically no-go decay. Ribosome profiling can reveal pausing, but proving NGD requires RNA-end mapping, decay-factor dependence, and rescue experiments.
> - Reporter assays require endogenous validation. A reporter isolates one feature but may lack native chromatin history, mRNP composition, splicing context, or dosage; endogenous transcript data are needed for broad claims.

## Biological Contexts Across Systems

Budding yeast remains a central model for defining deadenylation-dependent decapping and Xrn1-mediated decay. Mammalian systems add exon junction complex-linked NMD, complex tissue-specific regulation, immune and disease contexts, and abundant alternative splicing. Plants share many decay principles but have distinctive small RNA, uridylation, and developmental contexts. Viral infection tests host decay pathways because viral RNAs must be translated, replicated, and protected from host surveillance while sometimes manipulating decay machinery.

Developmental and cell-type contexts matter. Early embryos, neurons, immune cells, and differentiating lineages often regulate mRNA stability as strongly as transcription. A transcript that is unstable in proliferating cells can be stored in oocytes or localized in neurons. Therefore chapter-level claims should specify organism, cell type, stress condition, and assay whenever possible.

> **Box 35.2. NMD as Quality Control and Regulation**
>
> NMD acts differently depending on the transcript and disease context, and can be simultaneously protective, harmful, or regulatory.
>
> - Inherited premature termination codons: NMD can protect by reducing production of truncated proteins that would otherwise be toxic. In the same disease, NMD can be harmful if the truncated mRNA could instead produce a partially functional protein, making NMD net damaging for loss-of-function alleles.
> - Poison-exon isoforms and autoregulation: Many splicing factors include poison exons in their own pre-mRNAs; inclusion of these exons creates NMD-sensitive isoforms that limit expression of the splicing factor, coupling alternative splicing to gene-expression homeostasis.
> - Long 3′ UTR transcripts: Some normal mRNAs with extended 3′ UTRs behave as NMD targets, meaning NMD functions as a regulatory pathway for ordinary gene expression, not only as an error-correction system.
> - Viral RNAs: Viral mRNAs can be NMD targets, but viruses evolve structures, proteins, and replication strategies that evade or manipulate NMD, making the outcome virus-specific and context-dependent.
> - Cancer: NMD can protect against tumor-suppressor loss by removing nonsense-mutant transcripts, but can also harm immune surveillance by degrading neoantigen-encoding mRNAs or shape stress tolerance and growth control through regulation of nonmutant transcripts.

## Technology, Clinical, and Engineering Links

Therapeutic and engineered RNAs depend on decay biology. Synthetic mRNAs used for vaccines or protein replacement are designed to balance translation, innate immune sensing, tail length, cap structure, nucleotide modifications, codon usage, and stability. Antisense oligonucleotides can alter splicing so that NMD degrades an unwanted isoform or avoids degradation of a useful isoform. Small interfering RNAs deliberately trigger Argonaute-mediated target decay. Cancer therapies that modulate NMD are conceptually attractive but risky because NMD has transcript- and context-dependent roles.

> **Box 35.3. Tail, Cap, and Ribosome as a Three-Feature Stability Grammar**
>
> The decay risk of a cytoplasmic mRNA is jointly determined by three biochemical states rather than any single feature.
>
> - Poly(A)-tail state: A long tail fully coated with PABP promotes translation initiation and shields the 3′ end. Tail shortening by deadenylases reduces PABP occupancy and can shift the mRNP toward decapping, exosome access, or storage.
> - Cap state: An intact 7-methylguanosine cap recruits initiation factors and blocks Xrn1. Decapping commits the mRNA to 5′-to-3′ exonucleolytic decay and is the point of no return in the canonical pathway.
> - Ribosome state: Actively translating ribosomes protect the coding region and support the closed-loop mRNP architecture. Ribosome stalling, collision, or nonstop translation signals a defect and licenses quality-control decay pathways.
> - All three features act together: an mRNA that is translationally repressed, has a shortened tail, and has an accessible decapping complex faces a much higher decay probability than one failing only a single criterion.
> - [Chapter 36](chapter1034.md) adds a fourth layer: sequence grammar and codon optimality interact with these three features to set the baseline decay rate for individual transcripts.

Reporter design also depends on decay logic. A reporter with an artificial 3′ UTR, nonphysiological codon usage, unusual cap, or plasmid-driven overexpression can measure a pathway component but may not mimic endogenous mRNA turnover. Engineering stable expression requires controlling both synthesis and degradation.

## Recent Consensus

Current consensus treats eukaryotic mRNA decay as a regulated network, not a single linear pathway. Deadenylation-dependent decapping remains a central route for many mRNAs, especially in yeast and many cytoplasmic contexts. Decapping is actively controlled by cofactors and RNP remodeling. NMD is a translation-termination surveillance and regulatory pathway with organism-specific rules. Ribosome-associated decay pathways connect mRNA surveillance to protein quality control. Endonucleolytic cleavage can bypass cap and tail removal when targeted by small RNAs, ribosome stalls, stress pathways, or surveillance systems. Translation, localization, stress, and cell state are not external modifiers; they are part of the substrate-definition system.

## Open Questions, Controversies, Deprecated Models, and Common Misconceptions

Open questions:

- What determines how individual transcripts choose among decay routes in real time? Many diagrams show a dominant pathway, but single transcripts may face competing risks: continued translation, storage, deadenylation, decapping, endonucleolytic cleavage, NMD, NGD, or exosome-mediated decay.
- How can RNA decay be predicted quantitatively? Tail length, codon optimality, translation efficiency, RNA-binding motifs, and UTR length all influence stability, but no single feature predicts decay rate across all transcripts and cell types. [Chapter 36](chapter1034.md) treats sequence-directed stability models and their limits.

Common misconceptions:

- "NMD eliminates all mRNAs containing premature stops." NMD efficiency varies by termination context, cell type, gene architecture, translation, and surveillance-factor abundance. Some premature-termination-codon transcripts escape NMD, and some normal transcripts are NMD targets.
- "P-bodies or stress granules are decay sites." Granule localization can correlate with repression or decay, but direct decay requires evidence of RNA loss, cleavage, decapping, or factor-dependent turnover.
- "Low RNA abundance means rapid decay." Low abundance can reflect weak transcription, cell-type restriction, technical dropout, nuclear retention, or rapid decay. Decay claims require kinetic or mechanistic evidence.