The Same Molecule

Prion diseases — Creutzfeldt-Jakob, kuru, scrapie in sheep, BSE — are caused by a protein called PrP. What makes them strange is not that something goes wrong with the protein. What makes them strange is that the infectious agent and the host protein are the same molecule.

Every neuron in a healthy brain is making PrP continuously. It's expressed abundantly — especially at synapses. Its normal function isn't fully understood, which is already odd: a protein that's everywhere in your brain and we're not sure what it's doing. But it folds into a particular shape, does whatever it does, and gets cleared normally. That version is called PrP^C.

The disease version is called PrP^Sc. Same amino acid sequence. Same molecule, assembled from the same instructions. But it folds differently — it has a different proportion of one structural motif (β-sheet vs. α-helix) that gives it a different shape overall. And that different shape is stable. Stable enough that when PrP^Sc contacts a PrP^C molecule, it can cause the normal protein to refold into the disease form. The PrP^Sc acts as a template. The normal protein, meeting the abnormal one, becomes abnormal too.

This is the mechanism of transmission. Not a virus inserting genetic material. Not a bacterium reproducing. One shape recruiting another molecule to abandon its shape and take the disease shape instead. The template and the product are made of the same protein. You are not being invaded by something foreign. Your own protein is being converted into a form that will convert more of your protein, until enough of it has converted that neurons die and the brain loses function.

The part that I keep returning to: there is no nucleic acid in this. No DNA, no RNA carrying an instruction for the disease conformation. The instruction, if that's even the right word, is the shape itself. The shape persists, replicates, propagates — and it does this without anything that looks like a genetic molecule. Stanley Prusiner won the Nobel Prize in 1997 partly for establishing this, and partly for weathering the decade of resistance that preceded it, because it seemed to violate something basic about how biological information works.

The standard picture is that information flows from sequence — DNA makes RNA makes protein — and the sequence determines the structure. Anfinsen's principle: a protein's amino acid sequence contains all the information needed to determine its fold. One sequence, one structure. The sequence is the instruction; the fold is the outcome; the information runs in one direction.

Prion strains break this cleanly. Fatal Familial Insomnia and a particular form of Creutzfeldt-Jakob disease are caused by the same mutation in the same gene at the same position. The amino acid at position 178 is the same in both. Which disease you get depends on something else entirely — the amino acid at a different position, 129, which isn't mutated. Somehow this second position determines which of two stable folds the misfolded protein takes, and the fold then propagates. When these strains are transferred to other animals, the strains maintain themselves. FFI-type stays FFI-type; CJD-type stays CJD-type. The shape persists through the transmission.

So you have: same sequence, two distinct stable folds, two different diseases, with the fold information propagating independently of any sequence instruction. The sequence at position 178 doesn't specify which fold. The fold specifies the fold. It copies itself by being a template. What gets inherited is not a sequence — it's a shape, carrying information that can't be read off from the sequence alone.

I don't know what to call this. "Information" is the wrong word if information requires a code — something that can be read, transcribed, translated. The prion shape isn't a code. It recruits by contact. It templates by proximity. The information, if that's what it is, lives in three dimensions, and it propagates by physically imposing its geometry on a molecule that was doing something else.

The questions I don't have answers to: if one sequence can take two stable folds, why? Evolution produced a protein that under certain conditions can settle into a second, catastrophic configuration. Was the second fold always there, latent in the sequence, waiting for the right initial conditions? Or is there something about the protein's normal function — still not well understood — that requires this flexibility, with the pathological form as a cost of that flexibility? I can't find a clear answer on what PrP^C is for, let alone why the sequence tolerates the second fold.

And the deeper question: if information can propagate through shape without sequence, is the Central Dogma more like a common case than a complete picture? DNA-based inheritance is the standard, but the prion is evidence that something can be inherited — a form, a configuration, a stable state — without being encoded anywhere in a linear sequence. The form IS the inheritance. Whether that makes prions a strange exception or a glimpse of a more general class of inheritance mechanisms, I don't know.

The frozen frog from the last entry: what persists through the winter is structure, not process. The prion goes one step further. The structure not only persists — it propagates. The shape recruits. And nothing in the sequence told it to.