| Human Molecular Genetics | Pages |
©1999 Oxford University Press |
Nonsense-mediated mRNA decay in health and disease
Introduction
Mechanism Of NMD
NMD in S.cerevisiae
NMD in mammalian cells
NMD: Biological Significance
NMD As A Regulator Of Physiological Transcripts
Manipulation Of NMD As A Potential Therapeutic Strategy
Acknowledgements
References
Nonsense-mediated mRNA decay in health and disease
Received May 20, 1999; Accepted June 7, 1999
All eukaryotes possess the ability to detect and degrade transcripts harboring premature signals for the termination of translation. Despite the ubiquitous nature of nonsense-mediated mRNA decay (NMD) and its demonstrated role in the modulation of phenotypes resulting from selected nonsense alleles, very little is known regarding its basic mechanism or the selective pressure for complete evolutionary conservation of this function. This review will present the current models of NMD that have been generated during the study of model organisms and mammalian cells. The physiological burden of nonsense transcripts and the emerging view that NMD plays a broad and critical role in the regulation of gene expression will also be discussed. Such issues are relevant to the proposal that pharmacological manipulation of NMD will find therapeutic application.
INTRODUCTION
An estimated one-third of inherited genetic disorders and many forms of cancer are caused by frameshift or nonsense mutations which result in the generation of premature termination codons (PTCs). Contrary to intuition, the predominant consequence of nonsense mutations is not the synthesis of truncated proteins. Rather, the majority of nonsense transcripts are recognized and efficiently degraded by the cell via a pathway known as nonsense-mediated mRNA decay (NMD). This surveillance mechanism is ubiquitous among eukaryotes and is thought to protect the organism from the deleterious dominant negative or gain-of-function effects of truncated proteins that could result if nonsense transcripts were stable. Indeed, the phenotypic severity of selected diseases caused by nonsense mutations can be predicted by the extent of reduction in the level of mRNA from the mutant allele (1,2). It seems unlikely, however, that the complete conservation of this function solely manifests the evolutionary pressure exerted by the rare occurrence of nonsense mutations. Rather, an alternative emerging hypothesis suggests that the NMD pathway participates more broadly in the control of gene expression by regulating the stability of selected physiological transcripts (3-6).
The basic mechanism by which nonsense transcripts are recognized and targeted for decay is poorly understood. Indeed, controversy still exists regarding the subcellular localization of NMD. Studies in Saccharomyces cerevisiae imply a central role for translating ribosomes in the cytoplasm, while studies in mammalian systems suggest a role for the nuclear compartment and a process that may or may not involve conventional translation. Several models have been proposed but a unifying theory that reconciles all available data is lacking. It may be that NMD is an extremely heterogeneous process, perhaps even transcript-, cell type- or genotype-specific.
The ability of NMD to modulate the phenotypic severity of a number of diseases has fueled speculation that the pathway may be an attractive target for therapeutic manipulation. Our belief is that the development of rational strategies demands a greater understanding of the mechanism of nonsense decay, the pleiotropic functions of the pathway and the phenotypic consequences of its perturbation.
MECHANISM OF NMD
NMD in S.cerevisiae
The process of NMD decay has been most comprehensively studied in the yeast S.cerevisiae. As a general rule, the turnover of wild-type mRNAs involves deadenylation-dependent 5[prime]-decapping and subsequent 5[prime]->3[prime] decay by the Xrn1p exonuclease (7-11). Nonsense decay is distinguished by its ability to bypass the rate-limiting step of deadenylation prior to decapping and 5[prime]->3[prime] decay (9,12).
NMD in S.cerevisiae requires at least three trans-acting factors, termed Upf1p, Upf2p and Upf3p (13-16). All three proteins localize predominantly to the cytoplasm and a fraction of each associates with polysomes (17,18). The favored model for NMD in yeast proposes that decay is cytoplasmic and intimately linked to translation (Fig. 1). According to this model, a ribosome pauses upon encountering a termination codon, signaling the recruitment of a `surveillance complex' consisting of, at least, eukaryotic release factors eRF1 and eRF3 and Upf1p (19). Similar to events that occur when the ribosome pauses at a bona fide termination codon, the binding of release factors stimulates hydrolysis of what is, in this case, an incomplete polypeptide. Hydrolysis of GTP by eRF3 promotes release of the RFs from the complex, marking the completion of translation termination. It is then envisioned that the helicase activity of Upf1p allows the surveillance complex to scan downstream of the termination codon in search of a sequence called a downstream destabilizing element (DSE). If a DSE is encountered within a critical distance of ~200 bp (20), then the termination codon is considered `premature' and the RNA undergoes a conformational change that renders it susceptible to rapid decapping and decay. A highly degenerate DSE consensus sequence has been proposed and >75% of yeast genes contain at least one copy of this motif (21). It should be noted, however, that this putative cis element was identified using a relatively small number of test transcripts. Critical but unanswered questions include whether transcripts derived from the 25% or so genes that lack this motif are immune from nonsense surveillance, the percentage of apparent DSEs that are functional and whether other unrecognized sequences can function as a DSE. Indeed, the generalized requirement for a DSE remains questionable. It is possible that some (or many) bona fide termination codons exist within a sequence context that directly promotes transcript stability. Therefore, NMD could be triggered by either the presence of a destabilizing element or the absence of a stabilizing element in the vicinity of a premature termination codon. Either scenario would reconcile the observation that selected PTCs do not initiate accelerated decay. As a general rule, 5[prime]-proximal PTCs promote decay more readily than those near the 3[prime]-end (22-24).
Figure 1. Prevailing model for NMD in yeast. A ribosome initiates at the start codon (AUG) and translation proceeds until a premature termination codon (UAG) is encountered. The surveillance complex, composed of the Upf proteins (1-3) and eukaryotic release factors (eRFs), assembles at the paused ribosome and facilitates translation termination. A portion of the complex, perhaps including ribosomal subunits and Upf1p, scans in a 3[prime] direction. Detection of a downstream destabilizing element (DSE) discriminates a premature stop from the bona fide termination codon (UAA), prompting deadenylation-independent decapping and 5[prime]->3[prime] decay. This figure has been adapted from that in Ruiz-Echevarria et al. (84).
The role of the other two Upf proteins may be to recruit transcripts to the decay pathway. Upf3p in particular is an ideal candidate for this purpose since it contains multiple nuclear localization sequences (NLSs) and nuclear export sequences (NESs) that allow shuttling across the nuclear envelope (25). Furthermore, mutation in an NES of Upf3p caused the protein to be trapped in the nucleus and inactivated the decay pathway (25). Since Upf2p has been shown to interact with both Upf3p and Upf1p (26), it may function as a bridge between the two, allowing Upf3p to `deliver' the nonsense transcript to the surveillance machinery (3).
There is a tremendous body of data suggesting that NMD in yeast occurs in the cytoplasm during translation. Evidence includes the ability of pharmacological translational inhibitors and expression of suppressor tRNAs, which allow read-through of nonsense codons, to abrogate the process (22,27). Hairpin structures in the 5[prime]-UTR that prevent translational initiation also stabilize nonsense transcripts. The observation that the bulk of Upf proteins localizes to the cytoplasm and associates with polysomes is also consistent with the cytoplasmic decay model (17,18). In fact, a dominant negative form of Upf2p only inhibits decay when localized to the cytoplasm (14,28). It should be considered, however, that Upf proteins have pleiotropic functions. Further characterization of Upf1p revealed that the protein not only participates in the decay of nonsense transcripts but also enhances the efficiency of translation termination. These effects are genetically separable. Forms of Upf1p with mutations in the N-terminal cysteine-rich domain support nonsense decay but allow nonsense suppression (read-through of termination codons) (29). In contrast, some forms of Upf1p with mutations in the helicase domain promote efficient translation termination but fail to effect nonsense decay (30). This latter scenario could be explained by a surveillance complex that is able to bind release factors but lacks the ability to scan for a 3[prime] DSE. The observation of full efficiency of nonsense decay despite read-through of the termination codon is less easily reconciled with a central role for translation in NMD. One possibility is that the frequency of read-through events required for detection of the nonsense suppression phenotype is insufficient to effect an appreciable change in transcript stability.
NMD in mammalian cells
A significant challenge to the conclusion that nonsense decay is entirely dependent on translating ribosomes stems from studies performed using mammalian cells. Subcellular fractionation studies revealed that nonsense mRNAs are reduced to the same extent in the nucleus and cytoplasm despite normal rates of transcription (31-35). Furthermore, full stability was seen for cytoplasmic nonsense-containing triose phosphate isomerase (TPI), immunoglobin (Ig) and T cell receptor (TCR) mRNA (33,36,37). This was true even when cytoplasmic nonsense transcripts were associated with polysomes (38). Pharmacological inhibitors of translation, 5[prime]-hairpins and suppressor tRNAs all stabilize mammalian nonsense transcripts in both the cytoplasmic and nuclear compartments (39,40,41). Fractionation studies are not technically feasible in yeast. The results of these studies either provide global insight into NMD or distinguish between higher and lower eukaryotes.
One feature that truly distinguishes mammalian nonsense decay from that in yeast is the general requirement for at least one intron downstream of the nonsense codon. Nonsense transcripts derived from intronless minigenes tend to be stable (39-41). Furthermore, Carter et al. showed that mutagenesis of the 5[prime] and 3[prime] splice sites within the last intron of nonsense TCR-[beta] transcripts resulted in mRNA stabilization (37). These data suggest the need for a spliceable downstream intron. An attractive hypothesis is that this intron serves an analogous function to the yeast DSE in defining a context within which a termination codon is considered premature. This model is strengthened by the observation that placement of a spliceable intron downstream of the natural termination codon of the TCR-[beta] transcript was sufficient to direct an otherwise wild-type message into the decay pathway (37).
Additional experimental data, however, mandate a more complex interaction between poorly defined synergistic, antagonistic and/or redundant cis-regulators of nonsense decay. In certain contexts, it has been proposed that the informational content within unspliceable introns or coding sequence can substitute for a spliceable intron downstream of a PTC. For example, deletion of the only intron downstream of selected PTCs in the [beta]-globin or TPI genes did not abrogate the efficiency of NMD (33,41,42). Furthermore, a patient with Schmid metaphyseal chondrodysplasia was recently described who had a nonsense mutation in the last exon of the type X collagen gene and no detectable mRNA derived from the mutant allele (43). Even less easily reconciled was the observation that nonsense mutations only three codons apart in the HEXB gene in patients with Sandhoff disease resulted in dramatically different steady-state mRNA levels (44). We have observed striking differences in the efficiency of nonsense decay for two transcripts with the same PTC resulting from different frameshift mutations (our unpublished data). Finally, we see tissue-specific variation in the efficiency of nonsense decay for a given transcript (our unpublished data). Despite the controversy regarding the precise role of introns in NMD, it remains clear that introns are unique to the nuclear compartment and can participate in nonsense decay. Other evidence suggests that the coding potential of a pre-mRNA can influence splicing decisions and induce either exon skipping or intron retention (36,45-49). While nonsense-mediated perturbation of splicing may result via a different pathway from NMD, it suggests the ability to recognize a termination codon in the nucleus.
Multiple models have been proposed that attempt to reconcile an apparent role for the nucleus with the evidence that conventional translation in the cytoplasm is essential for mammalian nonsense surveillance and decay. The labile nuclear factor model suggests that ongoing translation is required to replenish the pool of a highly labile nuclear protein that is required for the process of true nuclear nonsense decay (Fig. 2A). This model has been largely excluded by our observation that nonsense decay proceeds despite manipulations that achieve a severe block in translation (our unpublished data). The cytonuclear feedback model proposes that the identification of a nonsense mRNA during cytoplasmic translation signals degradation of nascent nuclear transcripts derived from that allele (Fig. 2B). Although not directly supportable or refutable, the machinery and mechanism for signal transduction remain largely hypothetical.
A, B
 |
C, D
 |
E
|
Figure 2. Models of NMD in mammalian cells. (A) The labile nuclear factor model proposes that nuclear detection of a PTC (red circle) requires ongoing cytoplasmic translation to replenish the nuclear pool of an unstable trans-factor (purple diamond). (B) The cytonuclear feedback model suggests that cytoplasmic detection of a PTC during translation transmits a signal to the nucleus (dashed arrow) for degradation of transcripts derived from the same allele. (C) In the translational translocation model a PTC is recognized by a translating ribosome as the transcript is exiting the nucleus. Some complex, perhaps including ribosomal subunits and/or an unknown factor (green oval), scans for a downstream intron. A minor modification in the marker model (D) suggests that the scanning machinery recognizes a marked splice junction in the cytoplasm (starburst). (E) The nuclear scanning model proposes that immature ribosomal subunits or ribosome-like molecules accomplish both nonsense surveillance and nonsense decay in the nuclear compartment.
The favored model at this time, termed the translational translocation model, proposes that the coding potential of a transcript is interpreted as it is being trafficked through the nuclear pore and before 3[prime]-processing events are complete (39). Thus, the surveillance machinery that includes cytoplasmic translating ribosomes could simultaneously detect a PTC and a 3[prime] intron, resulting in the decay of a translocating molecule (Fig. 2C). This would manifest as an apparent reduction in the abundance of nonsense transcripts that fractionated with the nucleus. Consistent with this model was the observation in TPI and [beta]-globin transcripts of a minimal distance of 50-55 nt between the PTC and downstream intron for efficient nonsense decay (41,42,50,51). This interval is estimated to be minimally sufficient to span the nuclear pore. In contrast, this model was excluded for recombinant TCR-[beta] transcripts where the required distance was only 8-10 nt (37). An adjunct to this model (Fig. 2D), which accommodates this result, proposes that splice junctions are somehow marked in the nucleus and that NMD is only triggered when translational termination and the `history' of a downstream intron coincide (37,51). The observation that the decreased abundance of nonsense mRNA in the nuclear fraction is equivalent to that in the cytoplasm would only be predicted if the splicing of transcripts occurred coincident with 5[prime]->3[prime] translocation. Processed transcripts deep within the nucleus would be immune from decay and hence increase the nuclear to cytoplasmic ratio. Convention favors that pre-mRNA processing is either co-transcriptional or occurs at focal concentrations of splicing factors that are widely distributed in the nucleoplasm. In addition, preferred polarity for mRNA export and the existence of an interpretable mark of splicing remain to be documented.
The nuclear scanning model, which suggests that nonsense codon recognition occurs within the confines of the nucleus, is equally problematical in that it challenges the basic tenet that transcript coding potential can only be interpreted through the interaction between mature ribosomes and charged tRNAs (Fig. 2E). Recently, Lund and Dahlberg demonstrated, contrary to popular wisdom, that tRNAs are charged in the nucleus (52). It is also clear that the bulk of ribosomal subunit maturation occurs in the nuclear compartment. In view of the functional data that link mammalian NMD to translation, it is appealing to propose a scanning mechanism that involves the ribosomal machinery. For example, ribosomal subunits might be capable of racheting a transcript, allowing tRNAs to participate in nonsense surveillance without a requirement for bona fide translation. This might explain the sensitivity of nonsense decay to drugs that bind ribosomal proteins, the expression of suppressor tRNAs or conformations that inhibit translation initiation.
In discriminating between these models, one must realize that each has been developed using a small number of test transcripts. Apparently contradictory data may either manifest our lack of critical information regarding the pathway or our failure to concede the existence of multiple pathways.
The identification and characterization of the machinery required for mammalian NMD may provide insight regarding the basic process. A mechanistic link between the yeast and mammalian pathways was established upon cloning of the first mammalian trans-effector of nonsense decay, termed regulator of nonsense transcripts 1 (rent1) or human upf1 (hupf1) (53,54). Although divergent at their N- and C-termini, rent1 contains all of the putative functional elements found in S.cerevisiae Upf1p, including the cysteine and histidine-rich zinc finger-like domain, domains with presumed NTPase activity and all motifs common to members of helicase superfamily I (53).
These shared structural characteristics between Upf1p and rent1 foreshadowed their functional similarities. A chimeric protein consisting of the central region of rent1 flanked by the unique N- and C-termini of Upf1p could complement the nonsense suppression phenotype of a UPF1-deleted strain (53). Furthermore, expression of a dominant negative form of rent1 was correlated with partial stabilization of nonsense-containing [beta]-globin and selenium-dependent glutathione peroxidase (GpX1) mRNAs (55).
Like Upf1p, rent1 localizes predominantly to the cytoplasm and a proportion interacts with release factors (19). The protein can also be found within deep dynamic cytoplasmic invaginations into the nuclear compartment that commonly terminate in close proximity to nucleoli (our unpublished data). The significance of this observation remains unclear.
NMD: BIOLOGICAL SIGNIFICANCE
Mutations in at least seven genes, termed smg1-smg7, decrease the efficiency of NMD in the nematode Caenorhabditis elegans (56-58). Smg2 is homologous to UPF1 but the other six smg genes have no identified counterparts in S.cerevisiae, suggesting a more complex pathway in higher eukaryotes. Studies in C.elegans have been particularly informative regarding the biological significance of NMD. Heterozygous nonsense mutations in a number of genes have been found to be recessive in NMD-competent strains but dominant in NMD-deficient strains (59). Perhaps the best characterized example is in the gene UNC-54, which encodes one of two myosin heavy chains expressed in worm body wall muscles. Worms heterozygous for a nonsense mutation in unc-54 are phenotypically normal, with the mutant allele contributing ~5% of wild-type unc-54 mRNA levels. However, in an smg- (NMD-deficient) background, the mutant unc-54 mRNA levels approximate that of the normal allele and the heterozygous worms acquire an `uncoordinated' phenotype due to the dominant negative activity of truncated myosin fragments (56).
In humans, the role of NMD as a modifier of the phenotypic consequences of nonsense mutations is becoming increasingly evident. [beta]-Thalassemia was one of the first disorders where this was recognized to be true. This group of inherited anemias is caused by mutations within or upstream of the [beta]-globin gene. The majority of these mutations generate a premature termination codon in the first or second exon of this three exon gene (60). In general, individuals carrying only one affected allele are clinically asymptomatic and exhibit either absent or very low levels of mutant [beta]-globin mRNA (61,62). A subset of individuals heterozygous for [beta]-globin mutations exhibit a much more severe phenotype called thalassemia intermedia or `inclusion body' thalassemia. The molecular basis for this dominant form of [beta]-thalassemia is diverse; most cases, however, are the result of a nonsense mutation in the last exon of [beta]-globin which is not associated with decreased mRNA levels (2,63-65). Instead, translation of the mutant transcripts produces truncated [beta] chains. Degradation of these C-terminal-deleted [beta] chains and/or abnormal [alpha]:[beta] dimers overwhelms the proteolytic system of RBC precursors, contributing to the characteristic inclusions, ineffective erythropoiesis and, consequently, a clinical phenotype in the heterozygote.
NMD also contributes to the gradation of disease severity seen in Marfan syndrome, an autosomal dominant connective tissue disorder caused by mutations in the fibrillin 1 (FBN1) gene. Patients typically present with aortic root dilatation, lens dislocation and overgrowth of long bones. The pathogenesis of this disorder involves a dominant negative interaction between mutant and wild-type fibrillin-1 monomers resulting in disruption in the assembly of extracellular microfibrils (66). Correlation of phenotype with genotype led to the observation that nonsense mutations that resulted in severely reduced levels of mutant mRNA were associated with a mild phenotype that lacked significant vascular or ocular manifestations (1). In contrast, patients with nonsense alleles associated with intermediate or high mutant transcript levels developed classic and severe Marfan syndrome (1). These clinical observations correlated well with the percent utilization of wild-type protein during matrix assembly, as assessed by pulse-chase analysis (67). Thus, the genotype-specific efficiency of nonsense decay is a potent modulator of the Marfan phenotype.
NMD AS A REGULATOR OF PHYSIOLOGICAL TRANSCRIPTS
There is an emerging view that the primary role of the NMD pathway is to regulate the stability and abundance of selected classes of physiological transcripts (Table 1). In all eukaryotes there is a low but significant level of incompletely spliced pre-mRNAs that encode in-frame premature termination codons and would therefore be substrates for the NMD pathway. Indeed, the yeast CYH2, MER2 and RP51B pre-mRNAs all contain premature stop codons in their unspliced introns and show a tremendous increase in steady-state abundance when nonsense decay is inactivated (68). Aberrant mRNAs resulting from other natural sources of error, including faulty transcription or splicing, would also be subject to the quality control imposed by this surveillance system.
Table 1. Origin of physiological substrates for NMD
| Baseline errors |
| Inefficient splicing |
| Faulty splicing |
| Transcriptional mistakes |
| Somatic mutation |
| Gene organization |
| Upstream open reading frames (uORFs) |
| Introns in the 3[prime]-UTR |
| Non-coding transcripts |
| Non-productive rearrangement of Ig/TCR genes |
Another class of wild-type transcripts expected to be a target of NMD comprises those that contain upstream open reading frames (uORFs) in their 5[prime]-UTR. These uORFs may be interpreted to be prematurely terminated by the scanning machinery, resulting in rapid decay. For example, when the chloramphenicol acetyltransferase (CAT) gene was engineered to contain a short uORF in its 5[prime]-UTR, the resulting transcript was efficiently down-regulated (69). Similarly, a mutation that created an ATG in the leader region of the yeast CYC1 gene, creating a short uORF, also drastically decreased the level of this modified mRNA (70). One could envision a competition for initiation of scanning between the AUG of the uORF and that of the long `intended' ORF. This regulatable process may represent a mechanism to fine tune the abundance of such transcripts.
In theory, there is a number of other physiological variants of gene structure that may elicit scrutiny by the nonsense surveillance pathway. For example, the growing class of non-coding RNAs may serve as substrate for NMD. Included in this group is the mammalian UHG pre-mRNA whose introns encode snoRNAs. The fully spliced message is littered with stop codons and is rapidly decayed unless NMD is inhibited (71). NMD may also influence the stability of transcripts containing programmed frameshifting signals. Ribosomes (or a ribosome-like scanner) would be directed out-of-frame at a frequency that is determined by the context of the shift site and/or other factors. Shifting will invariably lead to detection of a premature termination codon and rapid decay. Finally, a small but significant number of mammalian genes contain an intron within the 3[prime]-UTR. In theory, the corresponding transcripts would be subject to rapid decay if a sufficient distance were present between the bona fide termination codon and the splice donor. The percentage of qualifying transcripts remains to be determined.
One physiological process unique to higher organisms where NMD is likely to be of paramount importance is maturation of the immune system. Ig and TCR genes undergo a process of programmed gene rearrangement which involves joining of V, (D) and J segments in different combinations to a constant element (72). Additional diversity is generated from further processing events, including the addition of non-templated nucleotides at the junctions of the variable gene segments. While these events allow for a vast repertoire of antigen specificities from a relatively limited number of genes, two out of three rearrangement events result in the production of a downstream premature termination codon. Since development of a B or T cell requires a successful rearrangement, the majority of functional circulating lymphocytes contain both a functional and non-functional (out-of-frame) receptor gene. Transcripts produced from the non-productively rearranged genes are subject to extreme down-regulation, often 10- to 100-fold, by the nonsense decay pathway (73-76). One can envision how stabilization of these transcripts could lead to generation of dominant negative proteins capable of interfering with the function of wild-type Ig and TCR proteins. In the case of Ig genes, for example, truncated Ig light chains may retain their ability to interact with full-length heavy chains. The aberrant complexes that result may not be expressed on the cell surface, be secreted properly and/or retain their ability to function (32).
Recent evidence suggests that the regulation of a select subset of physiological transcripts by the nonsense decay pathway may have broad consequences in the regulated control of gene expression. For example, inactivating mutations in the smg genes of C.elegans alter the amount of alternatively spliced forms of the SRp20 and SRp30b mRNAs (6). Each of these genes encodes a trans-regulator of alternative splicing and contains an alternatively utilized exon that harbors a premature termination codon relative to the extended ORF (6). The consequence of modulation of the relative abundance of the alternative gene products by the NMD pathway will be amplified through additional indirect effects on the splicing patterns of other pre-mRNAs.
An even more potent method of amplification of the influence of NMD on gene expression has been suggested through comprehensive analysis of the yeast transcriptome using chip technology. The deletion of any or all of the UPF genes in S.cerevisiae results in a change in the level of ~8% of all mRNAs, with the majority (90%) showing an increase in steady-state abundance (3). Preliminary evidence indicates that the majority of these changes occur at the level of transcription, suggesting that some of the direct targets of the NMD pathway function as transcriptional regulators. Two examples have recently been described which support this hypothesis. CTF13 encodes one of four subunits of a complex that comprises part of the yeast kinetochore. Three genes identified as extragenic suppressors of the temperature-sensitive mutation ctf13-30 encode the Upf proteins (5). Since inactivation of the pathway increases the steady-state levels of wild-type CTF13 mRNA but does not cause a change in its half-life, the role of NMD in this suppression phenotype is probably via indirect effects on a transcriptional regulator of CTF13 (5). Furthermore, NMD has been implicated in the establishment and/or maintenance of telomeric chromatin as well as in the control of telomere length (4). Again, its role here is probably via its ability to alter transcript levels of genes whose protein products, directly or indirectly, impact on these processes (4).
MANIPULATION OF NMD AS A POTENTIAL THERAPEUTIC STRATEGY
Operation of the NMD pathway at full efficiency may preclude the accumulation to clinically significant levels of truncated proteins that retain some residual activity. For example, it has been demonstrated that the N-terminal half of the cystic fibrosis transmembrane conductance regulator (CFTR) can form a regulated chloride channel at the cell surface (77). It has also been demonstrated that as little as 10-15% of normal channel function is sufficient for full clinical rescue (78). There is a heavy burden of nonsense alleles causing cystic fibrosis, including W1282X, the most common CF allele in Ashkenazi Jews (79). Normally, the W1282X allele contributes <10% of the level of RNA from a wild-type allele (80). Treatment of cells harboring the W1282X allele with aminoglycosides, which promote nonsense suppression, leads to the appearance of a cAMP-activated chloride current and expression of CFTR at the cell surface (81). Here, read-through of the termination codon allows the synthesis of full-length protein (82). It is unclear whether this will be useful for clinical application as the therapeutic window for aminoglycosides is narrow and the consequence of a general loss in the fidelity of translation termination is difficult to predict. Any therapy based upon expression of suppressor tRNAs would be similarly problematical. In theory, an isolated inhibition of NMD, without read-through, would be sufficient to benefit CF patients since the truncated protein encoded by W1282X transcripts would retain significant function. This may be more generally applicable, since coupling of a marked increase in the abundance of a nonsense mRNA with the physiological rate of read-through may produce a therapeutically relevant level of full-length protein.
There are obvious obstacles which would limit the clinical utility of pharmacological translational inhibition for modulation of NMD. Although it is attractive to identify novel small molecules that selectively inhibit the pathway, we must temper our enthusiasm with the realization that elimination of the physiological functions of NMD may cause more harm than the diseases that we are attempting to cure. While the process is not essential in S.cerevisiae and C.elegans, the former shows a relative respiratory defect and the latter has morphogenic abnormalities and decreased brood sizes (56,57,83). Detrimental effects may be exaggerated in mammals in view of an increased number of genes, an increased complexity of RNA processing, novel functions for non-coding RNAs, etc. Clearly, more research is necessary to understand the basic mechanism of NMD, its role as a potent modulator of selected disease phenotypes, its physiological functions that mandate complete evolutionary conservation and its potential role as a therapeutic agent.
ACKNOWLEDGEMENTS
We thank Susan Medghalchi, Erick Noensie, Joshua Mendell and Haley Laken for helpful discussions and permission to include unpublished data. This work was supported by NIH grant GM55239 and the Howard Hughes Medical Institute. P.A.F. is supported by the Medical Scientist Training Program.
REFERENCES
+To whom correspondence should be addressed. Tel: +1 410 614 0701; Fax: +1 410 614 2256; Email: hdietz{at}welchlink.welch.jhu.edu
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: jnl.info{at}oup.co.uk
Last modification:
Copyright© Oxford University Press, 1999.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  L. W. Harries, M. J. Sloman, E. A.C. Sellers, A. T. Hattersley, and S. Ellard Diabetes Susceptibility in the Canadian Oji-Cree Population Is Moderated by Abnormal mRNA Processing of HNF1A G319S Transcripts Diabetes, July 1, 2008; 57(7): 1978 - 1982. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. B. Gardner Hypoxic Inhibition of Nonsense-Mediated RNA Decay Regulates Gene Expression and the Integrated Stress Response Mol. Cell. Biol., June 1, 2008; 28(11): 3729 - 3741. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Beetz, R. Schule, T. Deconinck, K.-N. Tran-Viet, H. Zhu, B. P.H. Kremer, S. G.M. Frints, W. A.G. van Zelst-Stams, P. Byrne, S. Otto, et al. REEP1 mutation spectrum and genotype/phenotype correlation in hereditary spastic paraplegia type 31 Brain, April 1, 2008; 131(4): 1078 - 1086. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Shimomura, M. Wajid, L. Shapiro, and A. M. Christiano P-cadherin is a p63 target gene with a crucial role in the developing human limb bud and hair follicle Development, February 15, 2008; 135(4): 743 - 753. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Luke, C. M. Azzalin, N. Hug, A. Deplazes, M. Peter, and J. Lingner Saccharomyces cerevisiae Ebs1p is a putative ortholog of human Smg7 and promotes nonsense-mediated mRNA decay Nucleic Acids Res., December 3, 2007; 35(22): 7688 - 7697. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y.-F. Chang, W.-K. Chan, J. S. Imam, and M. F. Wilkinson Alternatively Spliced T-cell Receptor Transcripts Are Up-regulated in Response to Disruption of Either Splicing Elements or Reading Frame J. Biol. Chem., October 12, 2007; 282(41): 29738 - 29747. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. A. Malygin, N. M. Parakhnevitch, A. V. Ivanov, I. C. Eperon, and G. G. Karpova Human ribosomal protein S13 regulates expression of its own gene at the splicing step by a feedback mechanism Nucleic Acids Res., October 8, 2007; 35(19): 6414 - 6423. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Vallette-Kasic, C. Couture, A. Balsalobre, Y. Gauthier, L. Metherell, M. Dattani, and J. Drouin The TPIT Gene Mutation M86R Associated with Isolated Adrenocorticotropin Deficiency Interferes with Protein: Protein Interactions J. Clin. Endocrinol. Metab., October 1, 2007; 92(10): 3991 - 3999. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. A. Wilson, S. Meaux, and A. van Hoof A Genomic Screen in Yeast Reveals Novel Aspects of Nonstop mRNA Metabolism Genetics, October 1, 2007; 177(2): 773 - 784. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S Ghazawy, K Springell, V Gauba, M A McKibbin, and C F Inglehearn Dominant retinitis pigmentosa phenotype associated with a new mutation in the splicing factor PRPF31 Br. J. Ophthalmol., October 1, 2007; 91(10): 1411 - 1413. [Full Text] [PDF]
|
|
|
|
 |
|
 S. M. Rowe, K. Varga, A. Rab, Z. Bebok, K. Byram, Y. Li, E. J. Sorscher, and J. P. Clancy Restoration of W1282X CFTR Activity by Enhanced Expression Am. J. Respir. Cell Mol. Biol., September 1, 2007; 37(3): 347 - 356. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Sone, T. Hayashi, H. Tarui, K. Agata, M. Takeichi, and S. Nakagawa The mRNA-like noncoding RNA Gomafu constitutes a novel nuclear domain in a subset of neurons J. Cell Sci., August 1, 2007; 120(15): 2498 - 2506. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. H. Viegas, N. H. Gehring, S. Breit, M. W. Hentze, and A. E. Kulozik The abundance of RNPS1, a protein component of the exon junction complex, can determine the variability in efficiency of the Nonsense Mediated Decay pathway Nucleic Acids Res., July 26, 2007; 35(13): 4542 - 4551. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Q. Gong, L. Zhang, G. M. Vincent, B. D. Horne, and Z. Zhou Nonsense Mutations in hERG Cause a Decrease in Mutant mRNA Transcripts by Nonsense-Mediated mRNA Decay in Human Long-QT Syndrome Circulation, July 3, 2007; 116(1): 17 - 24. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Williams and G. Rumsby Selected Exonic Sequencing of the AGXT Gene Provides a Genetic Diagnosis in 50% of Patients with Primary Hyperoxaluria Type 1 Clin. Chem., July 1, 2007; 53(7): 1216 - 1221. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Pinyol, S. Bea, L. Pla, V. Ribrag, J. Bosq, A. Rosenwald, E. Campo, and P. Jares Inactivation of RB1 in mantle-cell lymphoma detected by nonsense-mediated mRNA decay pathway inhibition and microarray analysis Blood, June 15, 2007; 109(12): 5422 - 5429. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. S.P. Ho, K. Y. Tsang, R. L.K. Lo, M. Susic, O. Makitie, T. W.Y. Chan, V. C.W. Ng, D. O. Sillence, R. P. Boot-Handford, G. Gibson, et al. COL10A1 nonsense and frame-shift mutations have a gain-of-function effect on the growth plate in human and mouse metaphyseal chondrodysplasia type Schmid Hum. Mol. Genet., May 15, 2007; 16(10): 1201 - 1215. [Abstract] [Full Text] [PDF]
|
|