Little more than twenty years after the structure of DNA had been elucidated, and the science of molecular biology was beginning to come into its own, the discovery of intervening sequences in genes – “introns” took biologists by surprise. It had initially been assumed that, since DNA coded for protein (via RNA), then all DNA coded for protein, in the more or less discrete units called “genes”. That there was a substantial proportion that did not appear to code for anything, and which moreover broke up the coding sequences of genes, was something of an embarrassment to the elegant explanations of life that were being put forward. It was as though, in the middle of every paragraph of a book, a random sentence (or more) of gibberish appeared.

The first obvious question to ask is: “If the gene has what looks like a load of rubbish in it, how can you still get a functional protein out of it?” To make a protein from a gene, you need to do two things. First of all, an enzyme “reads” along the DNA and “transcribes” it into an RNA strand. This transcript then attaches to a complicated little structure called a ribosome, which “translates” it by reading along the RNA strand and spooling out a string of the appropriate amino acids in the order that they are coded for, joined together to make the protein. The introns are read through by the enzyme that makes the RNA message, but then this transcript is “spliced” so that the introns are all removed (literally the non-coding bits are chopped out, and the ends of the coding bits are glued back together). This explained how the genome could cope with introns, but not why they were there in the first place.

Explanations for the phenomena of introns were immediately sought. Critically, because of the assumption that a gene is synonymous with a protein, introns were assumed to be non-functional, because they did not code for protein. Due to this assumed non-functionality, scientists had to find theories which explained the presence of introns due to other factors, in both an historical and evolutionary sense. Introns were found to be present in all eukaryotes (organisms which package their DNA with protein in a “nucleus”, from the single-celled protista, such as the familiar amoeba, to multicellular mammals such as ourselves). They were also found in some bacteria – the eubacteria, but not in the majority of prokaryotes (all of which are single-celled organisms that do not have a nucleus). From this information, only two broad conclusions were possible. Prokaryotes are a far more ancient evolutionary group than eukaryotes. So, either the introns had been a late evolutionary invention, appearing only in the more recent lineages of eubacteria and eukaryotes, or, they had evolved early on, but had subsequently been lost from most of the bacterial kingdoms.

Thus, even early on, these two major contending theories arose; the “Introns early” and the “Introns late” models. In 1978, Gilbert suggested that introns had evolved later on, because, by splitting up a gene, they allowed the shuffling of the protein-coding sequences (“exons”), which could both speed up evolution by allowing new arrangements by recombination, and increase genetic complexity by differential splicing. The hypothesis was expanded upon by Blake, who suggested that the exons could code for discrete “domains” in the protein. Proteins are known to be composed of such domains, which are basically just units or modules with a similar structure that can be stuck together to make your protein. Introns, then, essentially gave you more protein for your gene. If they are spliced out such that the peptide strings that make up your protein are stuck together in different orders (as is known to happen quite frequently), then you can get slightly different proteins from the same gene. Buy one get two free, but only if you first invest in introns to break up your coding domains. This meant that introns had to be a later evolutionary innovation – something that was inserted into an already existing gene, rather than being there in the first place.

Darnell and Doolittle argued for the opposite viewpoint. They reasoned that there was not a lot of point in breaking up a perfectly nice coding sequence for a long-term selective advantage, when there would be no short-term benefits. (Remember that if, as Gilbert suggested, introns allowed exon shuffling, this process would necessarily have evolved over a reasonable period of time, and could not be expected to be anything other than a nuisance when a gene is first interrupted by non-coding DNA). If the introns did not evolve later, then it followed that they were there in the first place.

The “Introns early” model was largely favoured to begin with, as indicated by the prevailing evidence at the time (somewhat biochemical stuff I thoroughly intend to avoid going into detail about), and because it was not realised that there are different types of introns. Now it is known that there are, broadly speaking, two classes of intron; the “self-splicing” introns (types I, II and III) and the “spliceosomal processing” introns. The former are found in eubacteria and eukaryotes, the latter in eukaryotes. The essential difference is that the self-splicing introns, as their name suggests, are able to catalyse their own excision from the RNA transcript, but the latter require a specialised organelle called the spliceosome to chop them out. This distinction sounds pedantic but is important – it means that self-splicing introns could be present in prokaryotes, but spliceosomal introns can only be present in eukaryotes because only these cells have organelles (organelles are little structures bound by membranes that have specific functions; e.g the nucleus for packaging DNA and controlling its transcription etc., the mitochondria for the chemical processes that release energy from glucose. The analogy is with the organs of the body).

The arguments have swung back and forward over the years but opinion has currently swung back strongly in favour of the introns-late model. (I will briefly present the arguments in favour of the introns-early model, but intend to focus on the introns-late as the most likely hypothesis and for which there is most supporting evidence). From a variety of experiments the following picture and arguments have been put forward. The introns-early model necessarily means that, if the introns were there in the first place, they should not interrupt domains within the gene, even while they interrupt the gene itself. In almost all cases examined, introns appear to have no respect for boundaries and hence this appears to be untrue. However, it does seem that some introns in ancient genes are found in the same places – they are conserved in different kingdoms, supporting the hypothesis. If introns were present in prokaryotes from the start, the hypothesis must also explain why most bacteria lost them. The standard argument is that prokaryotes are under huge selective pressure to streamline their genome and cannot tolerate any “extra” genetic material. This is certainly true – any genetic engineer will tell you that if you insert an extra gene or bit of DNA into a bacterial cell, unless it is useful to the bacteria itself (i.e. it is selectively advantageous), then, within a few generations, the insertion will degrade and ultimately is likely to be lost altogether. This is why artificially inserted genes are often coupled to genes for antibiotic resistance – not only does it provide a “marker” to test whether your gene of interest really went in, but also, by growing the bacteria on a medium that contains the antibiotic, you ensure that to survive the bacteria have to keep that gene. However, as far as introns are concerned, this powerful argument shoots itself in the foot. If the majority of bacteria used to have introns but then lost them due to the selective pressure against the energetic cost of carrying all that extra DNA around, then there would have been a period of 1-2 billion years when the bacteria had introns, and didn’t seem to mind them in the least, followed by a rapid loss when there is no apparent reason for them to do so after that huge amount of time. This is simply implausible.

Considering again the different types of introns, a plausible model of their evolution can be drawn. First, introns were originally self-splicing “selfish” DNA - parasitic elements that inserted into the genome of bacteria. Since they were self-splicing, they could survive even though the bacteria had no means of removing them from the RNA themselves. However, prokaryotes have no nucleus. This means that transcription and translation (DNA to RNA then RNA to protein) are coupled processes. In this scenario, it’s difficult to have introns without them messing up the protein you get out of it because you go straight from your RNA message to your protein without a break in which any processing you may need can occur. Eukaryotes are believed to have evolved when smaller prokaryotes infected larger ones, eventually leading to symbiosis and the evolution of organelles from these symbionts. Introns could survive because they could go into the DNA in these organelles, which are discrete compartments in the cell. In the eukaryotes, the nucleus physically separates transcription and translation. Hence your RNA transcript can be held in the nucleus and processed before it goes into the cytoplasm to the waiting ribosomes to be translated into protein. Thus introns can be spliced out and the selective pressure to get rid of them completely is enormously reduced because then even if your gene looks a mess, your protein is fine. Hence introns were able to spread throughout the eukaryotic genome. Spliceosomes evolve to efficiently remove the now-ubiquitous introns, and thus some introns no longer need to self-splice, and lose this ability, leading to the situation today of eukaryotic genomes literally filled up with introns.