protein design using computational methods

Paul Barton-Davis pauld at stowe.cs.washington.edu
Fri May 17 10:54:33 EST 1991


In article <1991May17.121858.12141 at murdoch.acc.Virginia.EDU> wrp at biochsn.acc.Virginia.EDU (William R. Pearson) writes:
>   [ stuff from me about overloading of sequence-structure
>	relationships ]
>
>	I do not believe that there are any "well-known cases" of proteins
>of very similar sequence (>50% identity) folding into different
>conformations. I would be very interested in evidence to the contrary.

OK, I dug out my old copy of EMBL Research Reports to see if it
mentioned the ones I was talking about. Look in the literature for
stuff on the TIM barrel (triose phosphate isomerase). Chris Sander
would be a good investigator name to spot.

>Often, when X-ray structure people mention very different structures,
>they are referring to the orientation of the side chains or loops, or
>perhaps a very high precision statement about exact location of the
>alpha-carbons.  

This is often true. *HOWEVER*, it is frequently these small
differences that are critical in distinguishing functionality. One
consequence of this is that if there is this imprecision in the
relationship between sequence and structure, then although from the
point of biological investigation, not much is lost, we do lose the
ability to have the kind of engineering capabilities dreamed of by the
nanotech folks.
		 
>	For "proteins,"  however, those that are similar enough to be
>considered homologous ALWAYS have the same 3D structure.

Again, I believe if you look up work on the TIM barrel, I think you
will find some examples that contradict this, though I'm not sure to
what extent.

As a philosophical aside (I've been waiting to get this off my chest
for years, having been amongst computer geeks and not biochemists), I
do think there are some important reasons why we don't understand
protein folding and sequence/structure/function relationships in
general.  The primary one is that proteins and their interactions are
not too far above a level where quantum effects are still significant.
There are a few papers around in JTB on things like electron
tunnelling in proteins, and no doubt other effects at this scale exist
also. The interactions between a nascent polypeptide (boy, this
bio-jargon is more fun than CS), itself, ribosomes, translational
modification factors, the intracell environment, let alone substrates
in enzyme reactions, are fundamentally quantum mechanical (as are all
reactions). Researchers working on protein folding and protein-XXX
interactions (where XXX can be nucleic acids, lipids, water etc.)
generally limit themselves, often by practical necessity, to studying
proteins on a per-residue basis, with frequent dives down to the
side-chain atom level. I consider it unlikely that the interactions
are this crude. Proteins, with their magnificent conformational
contortions, offer nature the chance to create extremely subtle
variations in the physico-chemical environment for reactions. When we
talk of protein conformation, we normally talk of the relative
positions of residues or side chains. Isn't it more likely that the
functional aspects of protein conformation (both finished and during
folding) result from precise atom-atom positioning ? If this is true,
then it appears to me that we cannot understand protein folding until
we are in a position to understand these type of interactions.

This is not trying to imply that we can't make significant progress
toward understanding, at a gross level, the relationships between
structure and function. But just as in DNA, where it increasingly
appears that very subtle variations in atom positioning (e.g. tilt
angles between base pair planes) give rise to important functional
behaviour, the same is likely to be true of proteins. We are very
unlikely to be able to begin to engineer *new* proteins until we can
grasp how the complex quantum environment created in a protein
contributes to its function, and I would guess that that day is
several years away.


-- 
Paul Barton-Davis <pauld at cs.washington.edu> UW Computer Science Lab	 

"People cannot cooperate towards common goals if they are forced to
 compete with each other in order to guarantee their own survival."



More information about the Proteins mailing list