Let’s face it. Exons get all the glory. They’re the star players of the game, the quarterbacks, the offensive line. Meanwhile, introns and other intergenic regions play a more defensive strategy, are vitally important, yet rarely get their faces on sports cards, so to speak.
This whole notion of “junk DNA” still sours the scientist’s imagination even though it’s dawning on us that there is an elegant beauty and usefulness to these once-ignored segments. As an example of their importance from my own field of autism research, when you take a look at the variety of mutations which find association with the condition– even the big ones like CNTNAP2— very rarely do these mutations occur in exons but instead occur within intronic regions. We can’t yet say whether all these different mutations lead to alterations in gene expression or whether there is something more complex going on, but it certainly begs the question and suggests a more important role for introns than was once believed.
For those who follow this blog, you may be aware that I’m fascinated by transposable elements (TE). I can’t quite pinpoint what I find so fascinating except that 1) there’s so many TEs in the genome that it’s absolutely astounding, and 2) they largely seem to be ignored in the field of genetics– which I find rather baffling. At most they tend to be treated as an interesting phenomenon but are not foundational to broader genetics theory– even though the entire development of transgenic animal lines for research hinges on transposition. While I don’t expect every geneticist to dedicate his or her academic life to focusing on mobile elements, the fact that so many seem so uninterested is pretty sad. It’s tantamount to working in the field blinded to an entire range of effects. Just think how genetics theory may currently be closed off to alternative interpretations of data!
One thing which I find really interesting about TEs is that, by and large, they tend to be relegated to intronic regions. The table below is a few examples pulled from my own work, showing the four major categories of TEs in humans and the number of TEs in intronic regions of a given gene. A correlation between the categories, LINE and SINE, is to be somewhat expected since (autonomous) LINEs produce protein products which make it possible for (nonautonomous) SINEs to insert themselves into the DNA. But note that the LINEs, SINEs, LTRs, and DNA tranposons all correlate well with one another. To me it suggests there is something unique to these intronic regions that, when they attract one type of mobile element, they generally attract all of them.
At first glance, we may look at this tendency toward intronic insertion as something “adaptive”, i.e., if TEs targeted exonic, promoter, or enhancer regions with greater frequency, that would be maladaptive and those cells would more likely die off; therefore, those cells which allowed the insertion of TE content into the intronic regions, while possibly altering gene expression in some ways, were not as detrimentally affected and therefore lived to pass on this tendency to their progeny.
On some level, this may well be true. But for my part, I also wonder whether there isn’t something particular about the nature of intronic sequences which attracts transposition in the first place.
In previous blogs, I have stated that contrary to how most of us have been taught to imagine the genome (an analogue sequence), its 3D formation is vital to how the genome works. –In fact, I shouldn’t even say “vital” in the sense it is one part of a larger whole. DNA works the way it does, just as proteins do, in large part because of its 3D conformation. And as the term “conformation” implies, it is a dynamic form. It’s doesn’t remain in a complacent double-helical state but continually changes its superhelical twist, and parts of it jut out into alternative conformations like triple helices, cruciforms, and hairpins, in reaction to tension, binding partners, and the like. While the literature tentatively suggests it, I have suspicions that some of these hairpins and cruciforms may actually provide targets for transposons. Call it a hunch. Because such events, in relation to transposition and even daily events like transcription, are described in the literature over and over again. Too often to be coincidence. There’s something useful about these alternate forms and some day I’m sure we’ll know what it is.
Given my interest in alternative conformations, I came across an image in the Color Atlas of Genetics, 4th Ed. which struck me as fascinating. I think I had seen something like it way back when but until now it hadn’t really caught my attention. It was an electron micrograph of the hybridization of the original Ovalbumin DNA sequence and one of its RNA transcripts. From the image, you can see that the transcript has only hybridized with the exons of the DNA while the introns jut out and form a loop formation. The red line indicates the RNA being transcribed:
While the entire sequence is originally transcribed, the introns are spliced out post-transcriptionally. But there is growing evidence to suggest that intronic DNA segments, especially those containing palindromic and other repetitive sequences, are capable and perhaps prone to hairpin formation . The loop formation above is not the “hairpin” formation I’m referring to; nevertheless, this figure got me thinking and putting words to the ideas that’ve been swimming around my head for the last half year.
Loops. Atypical conformations. Maybe these non-double helical events, these hairpins, these random coils, are all targets for transposable elements. For what reasons I don’t know. Perhaps in these conformations, the base-pairing bonds are fewer and weaker than those in the double helix. Because of this, maybe they’re less stable and more easily broken. Thus, perhaps they are simply “weak spots” in the gene, more prone to insertions, point mutations, and copy number variants. –Is it also not a surprise then that both V(D)J recombination and certain transposition events actually utilize a hairpin structure themselves to create double-stranded breaks? 
It’s just a hunch. There’s a lot of tantalizing evidence out there right now, with no doubt more to come, and I genuinely look forward to its synthesis. Time will tell what functions these alternate conformations play. In any case, it will lend towards a more complex and 3-dimensional understanding of how the genome works and, when mutations do occur, why that is.
As an aside, I hope some day this concept of “random” mutation is disposed of. Who ever heard of anything truly random in nature anyways? Isn’t the very premise of science based on the assumption that everything within the natural world has an explanation and is therefore predictable? “Random mutation”, ha! Silly, Rabbit, randomness doesn’t belong in science. 🙄