We just finished having our minds blown by this two-part lecture by Susan Lindquist.

Part I:

Part II:

She weaves together prion misfolding and associated (mammalian) pathology with evidence about the (not-necessarily-pathological) effect of prion aggregate formation in yeast, focusing on the epigenetic nature of the trait in yeast.  She ultimately builds to a hypothesis that the prion protein (in yeast, at least) has been selected by evolution precisely for its ability to misfold and thereby to enable a flood of phenotypic variation in yeast in response to environmental stresses.

As I understand it, the prion in yeast is actually just one region of a protein, of which another important region is responsible for enforcing the observation of the stop codon during protein translation.   In yeast cells where the prions are soluble, this protein moves freely about the cell and causes translation to stop at a stop codon.  In yeast cells where a “misfolded” prion occurs spontaneously or is introduced, quickly all of the prions will form aggregates, and in so doing, they take the translation termination factor out of circulation.  This means that many stop codons will be read-through on translation, which in turn means that thousands of unexpressed regions in the yeast genome will suddenly be expressed.  Because these regions are normally unexpressed, they are free to accumulate a large number of mutations and so their sudden expression results in a vast variety in yeast phenotype.

Though she doesn’t go into it, I’m assuming that most of these new phenotypes caused by read-through are not very adaptive.  But the prion aggregation is triggered by environmental stresses– heat, oxidation– so it sounds as though this all happens in a scenario where the yeast has “nothing to lose.”  Suppose 1000 genetically distinct yeast were exposed to a potentially fatal stress, and they all experienced prion aggregation, and as  a result starting translating thousands of new proteins each.  There might be one yeast cell among the thousand which hits upon a phenotype that allows it to survive the stress and go on to reproduce.

Importantly, if you had another 1000 yeast which did not have the prion protein, they wouldn’t be able to respond to stress this way, and so they wouldn’t produce one organism that survives the stress.  So in this way, the prion protein could actually be adaptive.

Professor Lindquist does not tie this idea back into the possible function of prions in mammals.  Unlike in yeast, prion protein is not bound to a translation termination factor in mammals, and it’s also hard to imagine how the adaptation described for yeast would help mammals survive, since (as far as I understand) the aggregation of prions in mammals only occurs in a subset of tissues (say, the thalamus in FFI) and plus it’s not clear how stress on those tissues relates to macro stresses that the mammal encounters and must try to survive.

But if her hypothesis DOES apply to mammals, this would suggest that perhaps suppressing the expression of PrP entirely would be a viable way to treat FFI and other prion diseases.  Until now I’d been assuming that the fact that PrP is highly conserved across organisms must mean that it does something awfully important.  But if that important thing it does is allow organisms to adapt to stresses over a geologic time scale, then while it may be important to a species it is not important to any one individual, and so if a way can be found to suppress the expression of PrP, we’d be in good shape.

Another issue which Prof. Lindquist raises is the heat shock protein hsp104, which appears to be able to dissolve protein aggregates.  Do we mammals have this protein or something like it?  Is the structure of this protein understood and is this a potential treatment path?

A final question– at times Lindquist appears to use the term “prion” to refer generally to any proteinaceous infectious agent, i.e. any protein capable of converting other proteins to its own misfolded state, and not to “the” prion protein which in humans is called PRNP.  Can anyone clarify if this is the case?