X-RAY DIFFRACTION STUDIES OF PROTEIN STRUCTURE AND FUNCTION 
JohnKuriyan, Ph.D., Assistant Investigator 
Current research in Dr. Kuriyan's laboratory is 
aimed at obtaining atomic-level descriptions of how 
proteins carry out their specific functions. The em- 
phasis is on determination of three-dimensional 
structure by x-ray crystallography and on subse- 
quent computer simulation of protein structure and 
dynamics. Projects are under way in three broad 
areas: DNA-replication proteins, src-related onco- 
gene products, and redox enzymes (the last with 
support from the National Institutes of Health) . In 
addition, methodological work aimed at improving 
the x-ray structure refinement process is also being 
carried out, with support from the National Insti- 
tutes of Health. 
In collaboration with Dr. Michael O'Donnell 
(HHMI, Cornell University) and Rene Onrust, the /? 
subunit of the Escherichia coli DNA polymerase III 
complex (PolIII) has been crystallized and its three- 
dimensional structure determined to 2.5-A resolu- 
tion by x-ray crystallography. DNA polymerases are 
enzymes that duplicate the information content of 
DNA by catalyzing the template-directed polymer- 
ization of nucleic acids. Polymerases that are in- 
volved in chromosomal replication, such as PolIII, 
are distinguished by their high processivity; i.e., 
they can perform rapid replication (1 ,000 bases/s) 
of very long stretches of DNA without dissociation. 
This property is conferred upon PolIII by the ^ 
subunit, which acts to clamp the polymerase onto 
DNA. The /3 subunit is bound very tightly to DNA. 
Once assembled on circular duplex DNA, for exam- 
ple, it cannot be readily dissociated. It is, however, 
nonspecific in its interactions with DNA and has 
been shown to move along duplex DNA freely. 
How does the /? subunit manage to bind DNA with 
both tight (nondissociative) and loose (nonspe- 
cific) interactions? The x-ray structure reveals that 
two molecules of the /3 subunit are tightly associated 
to form a donut-shaped structure that can encircle 
DNA. This unprecedented molecular topology is sur- 
prisingly symmetric. Each monomer consists of 
three domains of identical chain topology. Each of 
these is roughly twofold symmetric, with an outer 
layer of two |8 sheets providing a scaffold that sup- 
ports two a helices. Replication of this motif around 
a circle results in a rigid molecule with 12a helices 
lining the inner surface of the ring and with six /3 
sheets forming the outer surface. 
The high symmetry of the ring-like structure is 
well suited to slide along cylindrically symmetric 
DNA. Structural analysis of the E. coli protein sug- 
gests that eukaryotic DNA polymerases have proces- 
sitivity factors with similar architecture. These pro- 
teins are known as proliferating cell nuclear 
antigens (PCNAs). Crystals of yeast PCNA that 
diffract to 3-A resolution have been obtained in 
collaboration with Dr. Peter Burghers (Washington 
University) . Structural analysis of PCNA and higher- 
resolution studies of the /3 subunit are in progress. 
The laboratory is collaborating with various 
groups to purify and crystallize src-related onco- 
gene products and their cellular equivalents. Initial 
efforts were focused on the SH2 domain of the Y-src 
tyrosine kinase, in a collaborative effort with the 
laboratories of Drs. Hidesaburo Hanafusa, David 
Cowburn, and David Baltimore (Rockefeller Univer- 
sity) and Dr. Marilyn Resh (Sloan-Kettering Insti- 
tute). SH2 domains are modular units that bind to 
phosphorylated tyrosine residues and are found in 
many proteins involved in signal transduction. Most 
proteins that contain these domains also contain 
other modules with various catalytic or binding ac- 
tivities, and the SH2 domains serve to localize these 
diverse signal transduction proteins at different sites 
of tyrosine phosphorylation, such as activated 
growth hormone receptors. 
Mutations in SH2 domains have been shown to 
affect the cellular transforming properties of various 
oncogenes, and there is great interest in designing 
specific inhibitors that may modulate these proper- 
ties. About 30 different SH2 domains have been 
identified, and sequence analysis suggests that they 
are likely to share a common molecular architec- 
ture. However, they do not show sequence similar- 
ity with any proteins of known three-dimensional 
structure. This has prevented the application of 
computer modeling methods to model and predict 
their structure. 
The \-src SH2 domain has been overexpressed, 
purified, and crystallized in the laboratory. High- 
resolution structures of the SH2 domain complexed 
with two phosphoryrosyl peptides have been deter- 
mined by x-ray crystallography. These structures 
have shown that the SH2 domain has a novel archi- 
tecture for binding peptides and have provided the 
first view of phosphotyrosine-protein interactions. 
The recognition of the phosphotyrosine group by 
the SH2 domain involves unusual interactions 
wherein positively charged groups of the protein 
interact with both the negatively charged phosphate 
476 
