Science Wiki

Exons are the nucleotide segments of a gene that are retained and joined together in the final, mature RNA produced from that gene after RNA splicing has removed the introns. The term refers both to the DNA sequence in the gene and to the corresponding sequence in the RNA transcript. The word was coined by Walter Gilbert in 1978 from "expressed region", in contrast to "intron" (an intervening region), a term introduced by Gilbert in 1978 after the discovery of split genes by Phillip Sharp and Richard Roberts in 1977. Exons are found in the genes of eukaryotes, in many viruses of eukaryotes, and, rarely, in the genes of archaea and bacteria.

In a protein-coding gene, the exons together contain the sequence that is translated by the ribosome into a protein, along with the untranslated regions at the ends of the mRNA. Exons are also found in many genes that encode non-coding RNAs, such as long non-coding RNAs, and in some tRNA genes, in which a small number of introns are spliced out. In the human genome, exons make up only about 1 to 2 percent of the sequence, whereas introns and the sequences between genes make up most of the rest.

Adenovirus R-Loop Hybridization

In 1977, the groups of Phillip Sharp and Richard J. Roberts showed, by hybridizing adenovirus mRNA to viral DNA and examining the products in the electron microscope, that the mRNA is complementary to several separate segments of the genome, with loops of unpaired DNA in between. This showed that genes are "split" into pieces, and that the primary transcript is cut and rejoined. The two received the 1993 Nobel Prize in Physiology or Medicine. Gilbert's terms "exon" and "intron" followed in 1978.

Splicing

In eukaryotes, RNA polymerase II transcribes a protein-coding gene into a primary transcript, the precursor mRNA (pre-mRNA), which contains both exons and introns. The 5′ end of the pre-mRNA is capped, and the 3′ end is cleaved and polyadenylated. The introns are removed, and the exons are joined, by the spliceosome, a large complex of five small nuclear RNAs (U1, U2, U4, U5, and U6) and many proteins, which together form small nuclear ribonucleoprotein particles (snRNPs). Splicing is not a "massive" or a purely mechanical cut and paste. It takes place in two transesterification steps, in which the 2′ hydroxyl of a branch-point adenosine in the intron attacks the 5′ splice site (which gives a free exon and a lariat intermediate), and the 3′ hydroxyl of the freed exon then attacks the 3′ splice site, which joins the exons and releases the intron as a lariat. The catalytic core of the spliceosome is formed largely by RNA (U2 and U6 snRNAs), and it is related in mechanism to the group II self-splicing introns.

The boundaries of exons are marked by sequences in the pre-mRNA. In the major class of introns (the U2-type), the intron starts with GU at the 5′ splice site and ends with AG at the 3′ splice site, with a branch point and a polypyrimidine tract upstream of the 3′ splice site. A minor class of introns (U12-type, about 0.5 percent of introns) is spliced by a distinct spliceosome and has different consensus sequences, and some of these begin with AU and end with AC. The statement that the GU and AG are "invariant" is therefore true for most, but not for all, introns. The sequences at the splice sites are only part of the information. In many genes, exonic and intronic splicing enhancers and silencers, which are bound by regulatory proteins, are needed to define the exon.

Exon definition

In vertebrates, exons are short (the median of human internal exons is about 120 to 150 nucleotides), and introns are long (the median is about 1,500 to 2,000 nucleotides, with some introns longer than 100 kilobases). In such genes, splice sites are thought to be recognized across the exon, in "exon definition", in which U1 snRNP at the downstream 5′ splice site and U2AF at the upstream 3′ splice site interact across the exon, and the pairs of exons are then brought together for splicing. In organisms with short introns, such as yeast and Drosophila in many genes, splice sites are recognized across the intron ("intron definition"). The mean number of exons per human gene is about 8 to 10, and a few genes have many more, such as the gene for titin, with about 360 exons, and the longest exon in the human genome is more than 10 kilobases.

Exon junction complex and nonsense-mediated decay

After splicing, the spliceosome deposits an exon junction complex (EJC) about 20 to 24 nucleotides upstream of each exon-exon junction. The EJC remains on the mRNA after export from the nucleus until the first (pioneer) round of translation, when the ribosome displaces it. It is involved in mRNA export, in localization, and in translation efficiency. It also makes possible nonsense-mediated decay (NMD): a stop codon that is located more than about 50 to 55 nucleotides upstream of the last exon-exon junction is treated as premature, because an EJC remains downstream of it when the ribosome terminates, and the mRNA is degraded. This "50 nucleotide rule" is why mutations that create premature stop codons in the last exon, or in the last 50 nucleotides of the penultimate exon, usually escape NMD and may produce truncated proteins. The EJC is therefore not a marker of "internal boundaries" only, and it does not initiate NMD by itself. NMD is triggered by the interaction of the termination complex with the EJC and with UPF proteins.

Coding Sequences and 5′/3′ Untranslated Regions

The statement that "exons contain only coding sequence" is a common misunderstanding, and so is the claim that the UTRs are separate from exons. In a protein-coding gene, the 5′ untranslated region (5′ UTR) and the 3′ UTR are generally parts of the first and last exons (and, in some genes, of other exons), because they are part of the mature mRNA.

• First exon. It contains the 5′ end of the mRNA, with the 5′ UTR, and, in many genes, the start codon (usually AUG). In many human genes, the 5′ UTR is spread over one or more non-coding exons, and the start codon is in a later exon, so that the first exon is entirely non-coding. The 5′ UTR affects the stability of the mRNA, the scanning of the ribosome, and the efficiency of translation, with the role of upstream open reading frames.
• Internal exons. They are in the middle of the transcript and are usually entirely coding, though they can contain UTR sequence in some genes. Because the genetic code is read in triplets, the exon boundaries can fall within a codon. The "phase" of an intron (0, 1, or 2) describes whether the intron lies between codons or after the first or the second nucleotide of a codon, and exons that are skipped without a shift of the reading frame are those whose length is a multiple of three.
• Last exon. It usually contains the stop codon and the 3′ UTR, including the signals for cleavage and polyadenylation (the AAUAAA signal in most mRNAs). The 3′ UTR contains elements that affect the stability, localization, and translation of the mRNA, including binding sites for microRNAs. In some genes, the 3′ UTR is spread over more than one exon, and alternative polyadenylation produces transcripts with different 3′ ends.

The poly(A) tail itself is added after cleavage of the RNA, and it is not encoded by an exon. The 5′ cap also is added to the transcript and is not encoded.

Mechanism and types

In alternative splicing, different combinations of exons are joined from the same pre-mRNA. The main types are the skipping of a cassette exon, the use of alternative 5′ or 3′ splice sites, mutually exclusive exons, the retention of an intron, and the use of alternative first or last exons (by alternative promoters or by alternative polyadenylation). The choice is determined by splicing regulators that bind to cis-acting elements, such as exonic splicing enhancers (ESEs) and silencers (ESSs), and the corresponding elements in introns. Serine/arginine-rich (SR) proteins usually bind enhancers and promote the inclusion of the exon. Heterogeneous nuclear ribonucleoproteins (hnRNPs) often bind silencers and promote exclusion, but this is not a general rule, since the effect depends on the position of the binding site. The transcription rate, the structure of chromatin, and signaling also influence the outcome.

For a gene with n independent cassette exons, each of which can be included or skipped, the number of possible isoforms is 2ⁿ. For example, with 10 such exons, there are 2¹⁰ = 1,024 combinations. Some genes have a very large number of possible isoforms, such as the Dscam gene of Drosophila, which has 4 clusters of mutually exclusive exons with 12, 48, 33, and 2 alternatives, which together give 12 × 48 × 33 × 2 = 38,016 possible isoforms, more than the number of genes in the genome.

Extent and functions

Alternative splicing of multi-exon genes is common. About 95 percent or more of human multi-exon genes produce more than one transcript, as shown by high-throughput sequencing. The number of protein-coding genes in humans is about 19,000 to 20,000, and the number of distinct proteins is larger because of alternative splicing, although the number of functionally distinct protein isoforms is much debated. The figure of more than 100,000 proteins is commonly cited, but the number of isoforms that are found in the cell at significant levels and that are translated into stable proteins is lower in proteomic studies, and many low-abundance transcripts may be errors of splicing or non-functional. An example of the functional effect is an exon that encodes a transmembrane domain. If it is included, the protein is membrane-bound, and if it is skipped, the protein can be secreted (as for the membrane-bound and soluble forms of the Fas receptor, or of some immunoglobulin heavy chains in B cells). The inclusion or exclusion of an exon can also change the binding of a protein to its partners, its localization, its enzymatic activity, or the stability of the mRNA, for example by introducing a premature stop codon that triggers NMD, which is a mechanism for the regulation of gene expression (regulated unproductive splicing and translation). The description that exon skipping leaves the protein "freely secreted into the extracellular matrix" is simplified, since the soluble form is released outside the cell and does not necessarily enter the matrix.

Splicing patterns differ among tissues, developmental stages, and species. They are especially complex in the nervous system. Splicing is also a target of disease. It is estimated that a large fraction of disease-causing point mutations affect splicing, either by changing the consensus sequences at the splice sites or by creating or removing enhancers and silencers, or by creating cryptic splice sites. Examples include spinal muscular atrophy (in which a C to T change in exon 7 of the SMN2 gene causes exon skipping, and the antisense oligonucleotide nusinersen increases inclusion of the exon), and some forms of Duchenne muscular dystrophy (in which exon-skipping therapies restore the reading frame), and cancers with mutations in splicing factor genes such as SF3B1 and U2AF1.

Exon shuffling

Exon shuffling, a proposal of Walter Gilbert in 1978, is the idea that new genes arise by the rearrangement and duplication of exons that encode functional units. Many proteins have a modular structure with domains that are encoded by one exon or a few exons, and many domains are found in different proteins, as in the epidermal growth factor-like and kringle domains of proteins in blood coagulation and fibrinolysis (such as tissue plasminogen activator) and the immunoglobulin domain. The shuffling of exons is made possible by the large size of introns, since recombination within introns (or by transposable elements such as LINE-1, which can carry along neighboring exons by transduction) can duplicate or move an exon without altering the coding sequence. The shuffled exons need to be in a compatible phase, and exons of "symmetric" length (a multiple of three), that begin and end in the same intron phase, are used more often in shuffling and alternative splicing. This supports the idea that exon shuffling is an important mechanism in the evolution of metazoan proteins, especially of those with multiple domains. The claim that it is "heavily responsible for the explosive radiation of multicellular life" is not supported as a general statement, and it is a hypothesis for some gene families. Other mechanisms, such as gene duplication, point mutation, and changes in regulation, also contribute.

Introns-early and introns-late

In the "introns-early" hypothesis, which was also proposed by Gilbert, introns were present in the ancestral genes of the earliest cells, and exons correspond to primitive protein modules, so that the modern proteins were assembled by the joining of exons. In the "introns-late" view, introns invaded genes later, from mobile group II introns of bacterial origin, which entered the eukaryotic ancestor with the mitochondrial endosymbiont, and the exon structure does not reflect the ancient modules. The evidence is mixed, and most data support a mostly late origin of introns, although the early origin of some is not excluded. The correspondence of exons and structural domains is weak in general, since many exons encode parts of domains, and many domains are encoded by several exons.

Genomic Coding Density and Exome Sequencing

The proportion of the genome that is exonic varies widely. In humans, protein-coding exons make up about 1.5 percent of the genome, and the whole set of exons, the exome, comprises about 30 to 40 megabases (including UTRs). In compact genomes, such as those of yeast and bacteria, the proportion that is coding is much higher (about 70 percent in yeast, and about 88 percent in Escherichia coli), and few genes have introns (about 5 percent in the yeast Saccharomyces cerevisiae). In humans, the mean exon is about 150 nucleotides, and the typical coding exon has a length in the range of 50 to 300 nucleotides. For a gene with n exons and a mean exon length L_e, the mature mRNA has a length of about n × L_e. For example, a gene with 10 exons of 150 nucleotides gives a 1,500 nucleotide transcript (before the poly(A) tail), which is much shorter than the gene itself, which may be 20 to 50 kilobases, so that only a few percent of the gene is represented in the mRNA.

The exome is the target of exome sequencing, a method in which the exons are captured by hybridization and sequenced. It is cheaper than whole-genome sequencing and is used for the diagnosis of rare genetic diseases, because a large part of the known disease-causing variants lies in coding exons or in the nearby splice sites. Exome sequencing does not detect variants in deep introns and in regulatory regions, and it has a lower uniformity of coverage than whole-genome sequencing.

Non-Canonical Splicing Substrates

• Non-coding RNA genes. Many long non-coding RNAs (lncRNAs) are spliced and have fewer exons than mRNAs (often two or three), and some have lower splicing efficiency. Some small nucleolar RNAs and microRNAs are located within introns of host genes.
• tRNA genes. A few tRNA genes have a single small intron, which is removed by a tRNA splicing endonuclease and a ligase, and not by the spliceosome.
• Trans-splicing. In some organisms, exons from separate pre-mRNA molecules are joined (trans-splicing), as in the addition of a 22 nucleotide spliced leader to the 5′ end of many mRNAs in trypanosomes and nematodes.
• Circular RNAs. Back-splicing, in which a downstream 5′ splice site joins an upstream 3′ splice site, produces circular RNAs made of one or more exons.
• Self-splicing introns. Group I and group II introns, which are found in some organelle and bacterial genes, can excise themselves without a spliceosome.

Are the 5′ and 3′ untranslated regions (UTRs) considered part of exons?

Yes. By definition, an exon is any segment of a primary RNA transcript that is retained in the mature RNA molecule after splicing removes the introns. Consequently, the 5′ untranslated region (5′ UTR) and 3′ untranslated region (3′ UTR) of a messenger RNA are integral parts of the terminal exons (and occasionally internal exons), even though they are not translated into amino acids by the ribosome.

What is intron phase and why does it matter for exon skipping?

Intron phase describes the position of an intron relative to the triplet reading frame: phase 0 introns fall directly between codons, phase 1 introns interrupt after the first nucleotide of a codon, and phase 2 introns interrupt after the second nucleotide. A cassette exon is "symmetric" if it is flanked by introns of the same phase ( meaning its nucleotide length is an exact multiple of three), allowing it to be skipped or duplicated without inducing a downstream frameshift.

How does the Exon Junction Complex (EJC) trigger nonsense-mediated mRNA decay?

During splicing, the spliceosome deposits an Exon Junction Complex (EJC) 20–24 nucleotides upstream of every exon-exon boundary. During the pioneer round of translation, the ribosome displaces all EJCs upstream of the canonical stop codon (which typically resides in the final exon). If a nonsense mutation creates a premature termination codon more than 50–55 nucleotides upstream of the final exon-exon junction, the remaining downstream EJC recruits UPF surveillance factors to degrade the defective mRNA.

• Intron
• RNA splicing
• Alternative splicing
• Spliceosome
• Exome
• Exon shuffling
• Nonsense-mediated decay