Proteins are interesting because in traditional chemistry terms they are very easy to characterise – just a long, specific sequence of 39 or so simple units – but their actual behavior is a matter of how that chain naturally tangles up at much larger scales, which is determined by their environment as much as the DNA that codes for them. It’s kind of like, a crochet hat is literally nothing more than a piece of yarn – if you pull on the loose end you can straighten it out – but understanding that piece of yarn tells you nothing about the hat, and the same piece of yarn could equally be a scarf or an amigurumi C-3PO doll, depending only on the history of what happened to it.
Except protein chains are kinked and curled in such a way that, when jangled around in the right context, they knot themselves into particular shapes. Where that context includes other existing proteins, it’s easy to see on a hand-waving level how one misfolded protein can cause others to conform themselves the wrong way, particularly where a chiral center is inverted. You can imagine a Bob Fosse dance routine where one dancer mistakenly crosses their left leg over the right instead of vice versa, forcing all the other dancers in the line to do the same, and then they just carry on with the choreo reversed, until at some point they’re supposed to cross with another line of dancers and it turns into a big pileup.
That is to say, it’s not mysterious in principle how “correct” proteins could infect each other with defective conformations; it’s just that no one has worked out how to reason adroitly about protein folding at the enormous scales involved.
