Eric Kool (what a name, I wonder if he has a brother named Joe) at Stanford University has created a clever hack on DNA where instead of storing the customary two bits per base pair, it can store three bits. Here, he inserts a benzene ring into the chemical structure of the nucleic acids and creates an “expanded” base pair set, thus increasing the set of base pairs from C,G,T, and A to include xC,xG,xT, and xA. So now, instead of being able to store just A-T/G-C pairs, a piece of DNA can now store xA-T, A-xT, xG-C, and G-xC combinations (x-x combinations and non x-x combinations are disallowed due to spacing design rules imposed by the rigidity of the deoxyribose backbone). It’s like StrataFlash for your cell nucleus. Of course, there are no polymerases in the cell that can handle replicating these, and there are no metabolic pathways to synthesize these nucleotides, but Rome wasn’t built in a day either.
Okay, okay, so this wasn’t a name that ware–it’s coming soon, I promise, and it’s a pretty interesting one too, I think–but when I read the article in Nature, I thought it was just too cool not to write a short post about it. The thought that something as evolved and taken for granted as DNA can be improved upon is pretty exciting; there’s apparently a lot more to explore out there! Presumably, there is some marked downside to xDNA, otherwise, evolution would have picked up on it…perhaps the metabolic overhead of creating and maintaining all of these extra base pairs wasn’t worth the overhead of getting better coding efficiency. Small viruses could probably benefit from more coding density, but there’s that nasty interoperability problem of xDNA with regular DNA. Then again, evolution tends towards local minima, and perhaps xDNA is in fact superior but chance never lined up to put all the right factors together in a single cell to create a sustainable xDNA line. I wonder if there is some alien lifeform out there (or perhaps a yet undiscovered species on this good planet) that uses the xDNA coding scheme.
Here’s the image from the Nature article, which gives you a better idea of how this stuff works:
Well, wait, you normally only get *1* bit per base pair, right, since it can only be A-T or C-G… so this would actually be a 2:1 improvement (four combinations = 2 bits).
I’m probably missing something, though.