Nonsense-mediated decay: linking a basic cellular process to human disease

Synthesis of mRNA in the Eukaryotes is a highly regulated step-wise process. During transcription, the nascent mRNA is suitably modified to remove introns (splicing) and to add the 5´ CAP (capping) and the 3´ polyA tail (polyadenylation). Capping and polyadenylation are important to maintain mRNA stability, transport to cytoplasm and translation efficiency. The multiple levels of regulation, however, also means that there is a possibility of error(s) occurring at each step. Eukaryotes have, therefore, evolved multiple ‘surveillance’ mechanisms, among which nonsense-mediated decay (NMD) is perhaps the best studied. NMD ensures destruction of mRNAs that harbor premature termination codons (PTCs); translation of such mRNAs would otherwise result in synthesis of truncated proteins that may have harmful consequences for the cell. Degradation of such mutated mRNAs leads to loss of function of the corresponding protein, often resulting in a disease condition. It has been estimated that approximately a third of all genetic disorders may be caused by NMD [1]. NMD is a precisely regulated process and appears to be conserved among different eukaryotic species with important distinct features in mammals.

How does the mammalian cell distinguish a PTC from the authentic termination codon? Initial studies revealed that incorporation of an intron downstream of the authentic termination codon could activate mRNA degradation [2]. In addition, intron-less cDNAs that are harboring a PTC were immune to degradation [3]. This seemed to indicate a possible link between nuclear splicing and NMD. Subsequent studies have revealed salient features of this complex process. During or immediately after splicing, a protein complex, known as the exon junction complex (EJC), is deposited by the spliceosome approximately 20–24 nucleotides upstream of each exon–exon junction [4]. In addition, nascent mRNA is bound by the nuclear cap binding complex CBP20–CBP80 (NuCBP), which directs translation of the immature mRNA, a process known as ‘pioneer round of translation’. If the translating ribosomes encounter a PTC (located in any exon except the last), they stall and recruit translation release factors, eRF2 and eRF3, which, with Upf1 and SMG-1 (a kinase that phosphorylates Upf1), form the SMG-1–Upf1–eRF2–eRF3 (SURF) complex. SMG-1 interacts with Upf2 and Upf3 (which are components of the EJC), resulting in activation of kinase activity of SMG-1. Upf1 phosphorylation by SMG-1 leads to the NMD cascade, resulting in degradation of the transcript. NMD is triggered only if the PTC is located at least 50–55 nucleotides upstream of the exon–exon junction, indicating, perhaps, that if the PTC is located closer, it might destabilize the SURF–EJC interaction due to stearic hindrance. In case of an authentic termination codon (or a PTC that is located in the terminal exon), there is no EJC located downstream of the PTC and, therefore, NMD is not triggered. Once the pioneer round of translation is complete (and all EJCs are removed by the translating ribosomes), NuCBP is replaced by eIF4E (the cytoplasmic CAP binding protein), which regulates bulk protein synthesis in the cytoplasm. Mature mRNAs are, therefore, not subjected to NMD.

In lower eukaryotes, either EJCs have not been identified (in yeast) or they appear not to play any role in NMD (in Drosophila). Rather, distance between the termination codon and the polyA tail appears to be an important factor that distinguishes a PTC from a normal termination codon. Human β-hexosaminidase, Rous sarcoma virus transcripts and mouse Ig-µ are examples of mammalian mRNAs that do not require an intron downstream of the PTC to be degraded by NMD [5–7]. This EJC-independent mRNA decay may be dependent on the physical distance between the PTC and the polyA tail, similar to lower eukaryotes [5]. The cytoplasmic polyA-binding protein (PABP)C1 appears to have an inhibitory effect on NMD in mammals [8]. In fact, if the polyA tail is brought in close proximity of the PTC, it may override the effect of the EJC [8]. The decision to trigger NMD appears to depend on the balance between the NMD-activating potential of EJCs and the NMD-inhibitory properties of PABPC1 [9].

The EJC is a multiprotein complex composed of approximately ten proteins, out of which at least four are core proteins and others may be transiently associated [4]. The core proteins include Y14 and Magoh, which form heterodimers, are part of the splicing machinery and are deposited on the mRNA by the spliceosome [4]. The third important core protein is eIF4AIII, which belongs to the family of RNA helicases and has been thought to be the major anchor protein for assembly of the EJC [10]. Other components of the EJC include splicing coactivators RNPS1 (also a core protein) and SRm160, the alternative splicing factor Pinin, mRNA export factors and Upf3b [11].

Upf1 is perhaps the most well studied of all proteins involved in NMD. It has RNA-dependent ATPase activity, as well as ATP-dependent helicase activity [12,13]. Unlike lower eukaryotes, Upf1 knockout in mammals is not viable, indicating perhaps that it may have roles in addition to NMD. Upf1 has been shown to be involved in mRNA degradation processes independent of NMD. It is responsible for degradation of mRNA molecules bound by Staufen 1 [14]. Upf1 specifically targets histone mRNA for degradation at the end of S phase, upon inhibition of DNA replication [15]. Interestingly, Upf1 is also involved in maintaining stability of DNA replication during S phase. Upf1-knockout mouse cells are not viable beyond the embryo stage [16] and knockdown of Upf1 in stable cell lines results in a block in cell cycle; a similar effect on knockdown of Upf2 is not observed [17]. Upf2 and Upf3 are both components of the EJC. Mammals (unlike lower eukaryotes) have two Upf3 genes, Upf3a (Upf3) and Upf3b (Upf3x) [18]. Similar to other EJC proteins, Upf3a and b are both predominantly nuclear proteins [18]. Upf2 has been proposed to be the adaptor that brings together Upf1 and Upf3 to elicit NMD [19,20]. In fact, if Upf3b is physically brought in the vicinity of a PTC, it itself can elicit NMD [21].

Each of the four SMGs have been shown to be essential for NMD. SMG-1 is a member of the PI-3 kinase family and can phosphorylate serine and threonine residues in target proteins; its role in the phosphorylation of Upf1 is essential for NMD [22]. Similar to Upf1, SMG-1 also plays a role in genomic stability by phosphorylating p53 [23] and by maintaining telomere stability [24]. SMG-5, -6 and -7 promote dephosphorylation of Upf1, although they do not have phosphatase activity. Rather, they function by recruiting phosphatases, such as the protein phosphatase 2A [25]. It is now known that a cycle of phosphorylation and dephosphorylation of Upf1, regulated by SMG proteins, is necessary for NMD [22]. Interestingly, tethering SMG-7 to any part of an mRNA induces its rapid decay, whereas other factors would induce decay only when tethered in the vicinity of a stop codon [26].

The culmination of interactions between EJC and the SURF complex, located on terminating ribosomes, is the decay of the mRNA. In lower eukaryotes, the decay may occur from either end of the mRNA. In fact, studies on NMD revealed the first evidence for mRNA degradation from the 5´ end in Drosophila, as well as in mammals. An endonucleolytic cleavage event was also shown to occur during mRNA degradation. SMG-6 has been shown to be the Upf1-dependent endonuclease in mammals and in Drosophila[27]. Recently, it was shown that the PNRC1 interacted with Upf-1 and Dcp1a (member of the decapping complex) and was probably responsible for recruitment of the mRNA decay machinery [28].

Clinical heterogeneity is a hallmark of several genetic disorders and understanding it is important for better patient management. Mutations affecting different amino acids have variable effects on protein structure and function and can, therefore, result in varying clinical symptoms. Similarly, the fact that efficiency of mRNA degradation through NMD is variable (depending on location of the PTC and on other factors) also results in a corresponding heterogeneity in clinical symptoms. In addition, this variation can even alter the inheritance pattern of genetic disorders. One of the first diseases where the role of NMD was identified is β-thalassemia, an autosomal recessive condition caused by biallelic inactivation of the β-globin gene. An individual exhibits disease symptoms if they inherit two mutated alleles (one each from either parent) and is unable to produce sufficient levels of functional β-globin polypeptide. A ‘carrier’ inherits a normal allele from one parent and a mutated allele from the other parent and is asymptomatic since the normal allele usually sustains sufficient levels of functional β-globin. However, in a carrier, if the mutation results in a PTC in the third (last) exon, the resultant mRNAs escape NMD, leading to synthesis of a truncated protein that causes aggregation of hemoglobin. The aggregated hemoglobin overloads the protein-degradation machinery of red blood cells, resulting in clinical symptoms of intermediate severity. This condition is known as thalassemia intermedia or dominant thalassemia [29]. Therefore, NMD may be said to have a protective role in β-thalassemia, since it protects individuals from deleterious dominant effects of truncated proteins (and, in general, protects heterozygous carriers from having a dominant disorder). The autosomal recessive form of myotonia congenita (Becker form) is caused by compound heterozygous mutations in the chloride channel gene (CLCN1), which may include a premature termination codon in an upstream exon. However, a rare autosomal dominant form (known as the Thomsen form) may be caused by a PTC in the last exon. The mutant transcripts escape NMD and the resultant protein behaves as a ‘poison’ polypeptide as it interferes with functioning of the protein synthesized from the normal copy of the gene [30]. Many tumor-suppressor genes, such as p53, BRCA1 and WT1, function as multimers or are part of a multiprotein complex, and truncated proteins may exert a dominant negative effect when NMD is unable to remove transcripts harboring PTCs [31].

Nonsense-mediated decay also affects the inheritance pattern of dominant disorders and may result in variants that follow a pseudo-recessive mode of inheritance. Marfan syndrome is caused by mutations in the Fibrilin-1 gene and disease manifestation occurs even in the heterozygous condition (i.e., one mutant allele is sufficient to cause the disease). However, if the mutation results in the generation of a PTC, the resulting mutant transcript is degraded by NMD, while transcripts arising from the wild-type allele continue to synthesis functional normal protein [32]. In such cases, the disease manifests with milder symptoms (and exhibits a recessive pattern of inheritance). Therefore, NMD can convert a dominantly inherited disorder into a pseudo-recessive form.

In all the aforementioned examples, the proteins involved either function as multimers or reside in multiprotein complexes. Therefore, even if one allele is mutated, the resultant mutant protein behaves as a poison polypeptide, leading to formation of nonfunctional multimers/complexes (dominant trait). Upstream PTC mutations usually result in a milder phenotype (or result in a recessive inheritance pattern), whereas if the PTC is located in the last exon (or the transcript escapes NMD through some other mechanism), symptoms are more severe (or result in a dominant inheritance pattern). It is to be kept in mind, however, that not all truncated proteins will be harmful. In fact, it has even been suggested that the more potentially harmful the nature of a truncated protein, the greater the extent of degradation caused by NMD [29,32].

In complete contrast to the beneficial role discussed previously, NMD has a major deleterious effect on cells by causing degradation of transcripts that could otherwise lead to production of fully or partially functional proteins. NMD has been shown to be directly responsible for several genetic disorders, and designing efficient modes of intervention provides an excellent opportunity for commercialization [33]. In cases where a single-base mutation converts a sense codon into a nonsense codon, the intervention strategy should inhibit translation termination at the PTC so that full-length functional protein can be synthesized. However, care should be taken that suppression does not occur at the authentic translation termination signal, nor should it affect the termination signals in other ‘normal’ genes in the cell. If the PTC is generated indirectly by a mutation that changes the reading frame, then, in addition to bypass of the PTC, the authentic reading frame has to be restored in order to yield functional protein.

Important diseases where NMD has been targeted as a mode of therapy include Duchenne muscular dystrophy (DMD) and cystic fibrosis (CF). Most PTC-generating mutations in the dystrophin gene that cause DMD are located in the 5´ end of the gene and lead to a severe form of DMD due to insufficient levels of the dystrophin protein. Mutations in the 3´ end of the dystrophin gene, however, cause a milder form of DMD (called the Becker form), since the mutant transcripts escape NMD [34]. Approximately a fourth of all DMD patients harbor a PTC-inducing mutation in exon 51. Antisense oligonucleotides, complimentary to sequences in exon 51 have been tested to rescue protein function by inducing exon skipping [35]. The success of this approach, however, depends on the extent to which the altered protein retains its function.

Cystic fibrosis is caused by mutations in the CFTR gene. Truncated variants of the CFTR protein have been shown to retain sufficient activity. In such cases, inhibition of NMD itself might alleviate disease symptoms without any need to trigger translational read-through of the PTC [36]. However, it was also shown that a full-length protein resulted in better recovery in patients compared with a truncated protein [37]. The most common approach for inhibiting NMD pursued by several laboratories is the use of aminoglycoside drugs that allow read-through of the PTC. Aminoglycosides perturb the codon–anticodon interaction, thereby leading to incorporation of an amino acid instead of termination of peptide synthesis. Therefore, the full-length protein is synthesized, albeit an incorrect amino acid may be included in the polypeptide. Aminoglycosides have been tried, with varying results, by several laboratories for treatment of CF [37,38]. Aminoglycoside drugs cannot be used in a prolonged manner as they have toxic side effects. Alternative drugs with fewer side effects are being tried for therapy of CF [39,40].

Conclusion

Mammals are distinct among eukaryotes in that they appear to have evolved a comprehensive mechanism to regulate gene expression through alternative splicing. It has been estimated that approximately a third of all alternatively spliced transcripts include a PTC. Understandably, therefore, NMD may play a pivotal role in mammals to regulate gene expression compared with lower eukaryotes. PTCs are generated through various normal processes in the cell, including translation from upstream ORFs, alternative splicing, aberrant transcripts from pseudogenes and transposons, transcripts containing codons for modified amino acids (such as selenocysteine), noncoding RNAs and transcripts from T- and B-cell receptor genes. NMD might have evolved to protect cells from deleterious consequences of truncated proteins. Therefore, one problem associated with suppression of translation termination using aminoglycosides is the suppression of natural NMD targets. It is expected that, in future years, many laboratories will focus on improving NMD-blocking therapies to minimize off-target effects. One major area of research will definitely focus on understanding the NMD process. Although rapid progress has been made in the past several years, we are far from understanding all the nuance of NMD. In our laboratory, we have identified PTC-generating mutations in several genetic disorders that appear to satisfy all ‘rules’ for NMD, yet the mutant transcript is not degraded. In the coming years, one can also envisage application of the beneficial role of NMD in the management of genetic disorders. Transcripts that harbor a PTC in their last exon often escape NMD leading to synthesis of truncated proteins with harmful consequences. Targeted activation of NMD in the affected tissue may be a way to manage clinical symptoms in patients who harbor dominant-negative PTC-generating mutations, such as in the β-globin gene, leading to thalassemia intermedia, and in tumor-suppressor genes, leading to cancer.

Financial & competing interests disclosure

This work was mainly supported by a core grant from the Department of Biotechnology, Ministry of Science and Technology, Government of India to the Centre for DNA Fingerprinting and Diagnostics, Hyderabad, India.The author has no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.