diaz@aecom.YU.EDU (Dizzy Dan Diaz) (11/11/87)
The following briefly summarizes some of the highlights of the recent
PROTEIN DOMAINS symposium held 2-3 November at the Waksman Institute for
Microbiology at Rutgers University.
SECOND DAY
R. Bruce Merrifield
Rockefeller University
"The Development of Hormone Agonists Through Chemical
Synthesis of Peptides"
Merrifield's group is interested in designing glucagon agonists which
bind the glucagon receptor but don't mediate the resultant biological
activities (e.g., increased blood glucose). Unfortunately, much of the
allotted time was spend discussing solid-phase synthesis, the development
of which bought the speaker a Nobel a few years back. He discussed new
methods of sidechain protection during synthesis and well as the use of
light-sensitive attachments to the solid resin, making the removal of
finished product easier.
Kurt Wuthrich
Institut fur Molekularbiologie und Biophysik, Switzerland
"Three-dimensional Protein Structures in Solution as Viewed
by Nuclear Magnetic Resonance"
This was one of my favorite talks. I had heard that it was now
possible, but I didn't know how well it worked. "It" refers to the
determination of solution structures of macromolecules using NMR.
Wuthrich discussed the three major kinds of NMR techniques used:
Proton-proton NMR uses H-H connectivity to determine interatomic
distances; COSY uses through-bond connectivity; NOESY (Nuclear
Overhauser Enhancement Spectroscopy) uses through-space connectivity.
A recent Science article reviews 2D NMR.
Using their technique, Wuthrich and colleagues were able to detect
errors in the published sequence of a protein they were working on. A
problem is that 2 A, which represents the sum of Van der Waals radii of
the atoms in question, is the present limit of resolution. The NMR and
crystal structure of the protein tendamistat (sp?) were compared. The
alpha-carbon skeleton structures were almost identical. Sidechain
conformations were the biggest difference between the two, especially
with sidechains near the surface of the protein. Right now the
technique is most useful for small proteins <= 10 kD.
Russell Doolitte
UCSD
"Protein Evolution"
A domain, Doolittle reminded us, used to be defined as an independently
folding module in a protein. The definition has been blurred as
molecular biologists have continued to bastardize terms from other areas
of biology to describe phenomena in their own domain :).
The determination of the function of protein domains is made puzzling by
the apparent ability to delete entire regions of proteins without known
biological effect. The gene for clotting factor VII has an intronless
region coding for ~1000 amino acid residues. When this huge region is
deleted, there is no effect on biological activity.
While many computer-armed molecular biologists (those rogues!) claim
great significance for their searches for sequence similarity (not
homology, Doolittle emphatically emphasizes), the way to determine if an
alignment is truly significant is to scramble the "similar" sequence and
recheck the alignment. Such searches for similarity can be predictive,
as was the finding that the E. coli uvrA protein had a region similar to
the Zn-binding fingers of certain eukaryotic gene regulatory factors.
When they checked out uvrA protein what did they find? 2 moles of bound
Zn per mol of protein.
As for the great similarity vs homology debate caused by massive abuse
of the second term, Doolittle says that is two proteins are >100 aa
long and >25% identical, then it would be strong evidence for homology.
Howard Nash
NIMH
"Biochemistry of Site-Specific Recombination:
Structure and Function of a Multiprotein Complex"
Nash studies the integration of lambda phage as a model of
recombination. The 20-25 nt att site in E. coli is identical to
a stretch on the lambda chromosome, the site of crossover.
Int protein binds to both sequences; it is a site-specific
topoisomerase. Integration host factor is a heterodimeric E. coli
protein which forms part of the integration complex, the intasome.
The sequence of IHF is similar to that of HU protein, a histone-like
protein in E. coli. The structure of HU has been determined and by
comparison, they have theorized that IHF binds DNA in the minor groove.
This result is difficult to rationalize on the basis of present models
of DNA-protein interactions. The problem is that if one assumes only
H-bonding for specific binding, it is difficult to see how the available
H-bond contacts in the minor groove could foster such specificity.
--
dn/dx Dept Molecular Biology diaz@aecom.yu.edu
Dan Diaz Albert Finkelstein College of Medicine