clarify protein folding problem

Tom Kelting milemgr at
Sat Jun 30 15:31:43 EST 2001

I’d like a clarification on the nature of the protein folding problem.

Perhaps the best thing is for me to state what my reading is of the problem,
from what I’ve read, and then an expert can, hopefully, correct anything
that is wrong in my thinking, and expand upon it further.

A protein molecule is a very complex structure, occurring, of course in
three dimensions.  The nucleic acid codons that occur in the cell, specify
the general recipe for each protein… the sort of information contained in
its in-line molecular formula (formulas of the form C6-H12-O6), but not its

Just as there may be left and right chirality in a simple molecule, there
are different possible shapes that a molecule with a specified, inline
molecular formula, may assume.  These different shapes derive from the fact
that, in most instances, there are a variety of bond angles permitted by the
actual physics of molecular bonding.  An atom or small polymer that is going
to attach to an evolving protein molecule that is being built up, might be
permitted to attach to any of several positions.

Some such attachments would lead to “dead ends” – terminations of the
synthesizing molecule-in-progress before its target molecular formula could
be fully satisfied.  This would be one kind of error.

Other errors might lead to a molecule with the abstract formula being
satisfied, but with an entirely different shape, than the “desired shape”.

Am I right so far?  I have said that (1) genetic information specifies
molecular formula, but not shape, and (2) the first aspect of the
folding/construction problem is to efficiently construct the molecule
without leading to dead ends, partial molecular constructs that have to be
discarded because they are misshapen and incapable of further movement
toward the desired shape.

Now, the further question is, if we assume that there are multiple possible
shapes that might be satisfactions of the nucleic codon-resident, abstract
molecular recipe for this protein, nonetheless only *one* shape (with
perhaps tiny variations) would be the shape that has the desired, needed
functionality, in the relatively global protein ecology of that organism
(e.g. the cell).

Take ordinary symmetry-chirality.  I have read that, in situ, the “right”
chirality is constructed a particularly high proportion of the time (unlike
random statistical synthesis in the laboratory… as with generic drugs where
chirality is an issue.)

Anyway, whether this chirality claim is true or not, the issue (3) is this:
if it is true that multiple structural geometries are possible, in a
particular case, for a particular protein, then all are equally “right” from
a chemical and physical standpoint.

But the properties, the functional powers, within the relatively global
context of the protein ecology of the organism, differ with shape.  From
that standpoint of the ORGANISM, only one shape is the “right” one.  How is
*this* shape constructed and selected, from among the alternate, physically
permissible shapes?

This to me is a whole different kind of protein folding problem.  I would
imagine it as if, say, one chirality of a certain molecule is needed in a
certain token of an organism, and the same molecule might be needed in the
complementary chirality, in a different context … say, in a different
species, (or even in a different token of the same species).  The point
would be that there is some contextual, mesoscale global “right answer” that
is context dependent.  During molecular synthesis, somehow, the structure
that is needed in the current context, must be selected from among the
complementary, equally permissible shapes.  I call this point (3).

So, what is the protein folding problem… more like (2), or like (3), or are
my assumptions all off base.

And, if I am getting it right so far, how does computer simulation
(intractable in polynomial time and so forth), solve either 2 OR 3?  (I
especially don’t see how it could solve 3).  Just *exactly* what is it that
is being simulated, when a computer attempts to model protein folding, in
the bioinformatics research on protein folding that we always see referred
to in the news?

Can someone steer me right, if I am off track?  Thanks for your help.

More information about the Bioforum mailing list