Andart

Prions R' Us

I remember inventing the prion for one of my childhood comics. It was, if I remember right, about a space expedition to some alien world. One astronaut ingested some colorful alien chemicals (are there no dull alien chemicals? not when you draw with crayons) and they combined to form some kind of self-replicating green-red geodesic sphere in his bloodstream. The other astronauts somehow drew it out with a syringe. Anyway, I recalled that even at this tender age (perhaps around 8 years) I thought the scientific premise was a bit shaky. I felt vindicated when I heard about prions several years later. Now I get the feeling we should expect the colorful alien chemicals everywhere.

Kausik Si, Eric Kandel, Susan Lindquist et al. suggest in two papers in Cell that prions may play a role in memory. An Aplysia protein ApCPEB is upregulated by short term facilitation and necessary for long-term facilitation. It also exhibits prion-like behavior in that it can convert ApCPEB (at least in yeast) into a new conformational state that can convert ApCPEB into the new state. The suggested function is that the prion state acts as a synaptic marker activating mRNA to do long-term changes.

If this fascinating possibility works out and plays a role in memory Kandel will likely get another Nobel prize. And we will have to care about the conformome beside the genome, proteome, kineome, transcriptome and the others.

There is a lot of elegance in having a self-reproducing prion state - it is resistant to noise, can persist even if no unchanged ApCPEB arrives for a while and activate mRNA at leisure. It helps explain the problem Upinder Bhalla discussed when he visited us at SANS: how can you get reliability when the number of proteins at a synaptic terminal is so low? By having the effect persist for a long time one can likely get a lot of reliability. However, it requires compartmentalization to retain specificity. If the prions were allowed to run around they would make the neuron non-specific. Which makes it likely that ApCPEB in the activated state should not be allowed to diffuse out of the spines, or that there is something eating it in the dendrites.

It suggests a lot of fun modeling. First the basic ApCPEB dynamics, especially how necessary it is to prevent spontaneous prionization - do we need a serotonin trigger? How fast can it be broken down?

Then the compartmentalisation mechanisms. As well as the interesting issue of what happens if these break down: do we get a prion disease, or some form of memory impairment? Having synaptic markers overload individual cells is likely not noticeable, but what if the compartmentalisation went overboard on a large scale? Cells would become non-Hebbian and likely develop a lot of LTP or LTD; sounds like a fairly distinct pathology that might be observable. Also, one would expect different genetic influences on the vulnerability for this.