Evolution and Protein Folds

Erich Schwarz schwarze at starbase1.caltech.edu
Fri Jun 24 19:43:23 EST 1994

In article <2ufnhv$bpl at umd5.umd.edu>, ram at mbisgi.umd.edu (Ram Samudrala)

> I am now confused as to what you mean by "acceptable".  Who decides
> what is acceptable?  Remember, there's no selection on the protein
> produced by the gene yet.  What's to prevent it from mutating a
> residue in the hydrophobic core to an ASP and thereby destabilising
> the fold by a few Cal? Given that there's no selection on this
> protein, it doesn't seem to matter if it folds up or not.  But it
> does, if it has to evolve to a new function.  Thus there is some
> mechanism that makes sure a compact-fold is preserved.
> In any case, are there any numbers for such a thing?  How do our
> current models explain the times of divergence of
> haemoglobin/myoglobin? 
> Oh, the protein is just a copy of another one.  So there is one gene,
> producing fully functional proteins.  The other gene is the one that
> is mutation.  So there is no selective pressure on one copy of the
> gene and it mutates at will in the hopes of finding a new function.
> Now on the way, I claim it cannot have mutations that will destroy the
> fold.  If it does, then it will a lot longer to get back on track.
> What is the mechanism that ensures that no mutations that destroy the
> fold (again, there's really no protein deficiency) are not selected
> for?

    As I understand the current view:

    1. Gene duplication at levels from the exon to the genome happens quite
a lot.  Ohno wrote a classic text on this circa 1970, and Nadeau has
reviewed the evidence for this at the subchromosomal level fairly recently
(Mouse Genome, vol. 1 or so.)  Repeated sub-genic motifs are abundant
(check out the Sequence Analysis Primer, edited by Devereaux and Gribskov,
chapter 3.)

    2. Gene deletion, and nonsense/missplicing mutations that knock out
genes, are also common but less remarked upon.  In humans we actually have
quite an inventory of spontaneous deletions (manifested as inherited
diseases); see _Mendelian Inheritance in Man_ by McKusick.  There is no
reason to think that the fairly abundant deletions we see in _Homo sapiens_
are somehow peculiar to our species.

    3. Given gene duplication producing a surfeit of excess genes, one can
expect *both* advantageous gain-of-function mutations *and* deleterious
misfolding mutations to occur.  The former will get fixed in the species at
a mathematically calculable rate (check out any text on
population/evolutionary genetics.)  The latter will either select against
themselves if they produce dominant negative phenotypes, or just tend to
get stochastically winnowed out by either deletions or missense/nonsense
mutations that silence the misfolding protein's gene.

    4. The "selective pressure" for elaboration of the genome is not
understood at all in detail.  But we can vaguely speculate that big,
complicated organisms benefit from the biochemical flexibility afforded
them by having multiple, subtly different versions of a protein.  In
mammals this seems to be primarily achieved by gene duplication and
divergence.  But this is hardly the only means; multiple splicing can also
work, and in _Drosophila_, seems to actually be a preferred mechanism of
evolutionary complexification.  For example, consider the fact that in
mammals, _Shaker_-type potassium channels are a multigene family, but in
_Drosophila_ a bouquet of _Sh._ channels with subtle differences is
produced by multiple splicing from one gene.

    5. Does that all make sense?

--Erich Schwarz
  schwarze.ccomail at starbase1.caltech.edu
  Division of Biology, Caltech

More information about the Mol-evol mailing list