Marcio Pie wrote:
> [...]
> >The prion system, like a mutation in any translation factor, may
> >interfere with many genes at once. What makes the prion system a very
> >special case is that a prion is autocatalytically produced from its
> >"normal" cellular conformer. The yeast prion [PSI+] is a
> > [...]
> >Within the given cell line, this "epigenetic" change is heritable. But
> >it can be reversed by counterselection.
>
> But isn't it the germ line heritability what really matters? Isn't this
> just a kind of molecular-level phenotypic plasticity? If it is, what is
> so revolutionary about it to be published in Nature?
The yeast [PSI+] state acts as a cytoplasmically inherited
nonsense-suppressor (read-through of a stop codon), whereas normal
nonsense-suppressors for a specific type of stop codon (amber, ochre,
opal in bacteria - I assume they also occur in yeast) are genomically
inherited mutations (of course, single-cell organisms like yeast are
their own "germ line"). Both types of suppressor state can be reversed
occasionally, but I don't know how the frequencies of prion reversion
and genomic suppressor backmutation compare, and therefore I can't tell
what "really matters" in yeast regarding evolution or phenotypic
plasticity. Furthermore, prions can be artificially "cured" (by
guanidine hydorchloride), therefore this prion reversion might also be
induced by some natural environmental conditions - which would not be
the case for genome mutations.
> [...]
> >Both the pseudogene and the prion systems, however, can only activate
> >something that already exists. If this is to be a new functionality, it
> >must have emerged by a mutational random walk in the cryptic state,
> >hidden from natural selection. Can we expect really novel
> >functionalities to emerge by such random processes extending over
> >several mutations, but without intermediate selection?
>
> I totally agree with you here. It seems to me that, over time, there should
> be an exponential decrease in the probability that this potentially new
> function could arise. Maybe this is the reason why the rate of gene
> duplications seems to be surprisingly high (at least to me) in some cases
> (sometimes in the order of 10^-3 per generation!!). Eventually this
> pseudogene should be deleted, otherwise genome size would grow
> exponentially, together with the accompanying costs. By the way, do you
> have any idea of how this overflow of new pseudogenes could be detected
> and removed from the genome?
Why do you expect a decrease in the probability that new functions could
arise? IF they arise at all by random-walk mutational sequences in
unused DNA (to be detected and put to use by means of prions or other
suppressor mutations, or backmutations in pseudogenes), I would expect
the frequency of their emergence to remain stable, on the average.
Gene duplications occur by unequal crossing-over during DNA replication,
which, for unused DNA, will result in DNA loss about as often as DNA
gain, on the average (as for required DNA, of course, loss would be
lethal). So, here too, I would expect some kind of dynamic equilibrium.
The average percentage of unused DNA will be the result of a balance
between cost and benefit. The cost may be mainly metabolic cost, the
benefit might include security against deleterious mutations by
duplication of function, cryptic backup systems for special environments
or stress, raw material for microevolutionary possibilities.
Peter Ruest
This archive was generated by hypermail 2b29 : Mon Oct 23 2000 - 11:38:20 EDT