With genomics getting so much media attention, researchers working
on proteins could be excused for grumbling that it is proteins,
after all, that actually do most of the work in the cell. Technology
has enabled an accelerating stream of genome sequences, and powerful
new methods like gene expression arrays have identified countless
genes involved in virtually every disease and condition under the
sun. The study of proteins has been advancing as well, albeit more
quietly, and just passed a milestone as the Protein Structure Initiative
(PSI, an NIGMS effort) completed its first 5-year phase and moved
into its second.
The goal of the PSI is basically to make it easier to figure out
the three-dimensional shapes of proteins, with the long-term goal
of being able to predict most protein structures from their DNA
sequences. A genetic mutation is, to use an analogy, like a spelling
change in one letter of a long word. To understand how that mutation
leads to disease, it helps to know how it changes a protein's structure.
How proteins work, put simply, depends on their three-dimensional
structure.
But it's more than a better understanding of biology that's at
stake. For example, NIGMS-funded researchers at Yale recently identified
the structural changes that enable bacteria to become resistant
to some antibiotics. Researchers knew that a mutation in a ribosome
gene led to the resistance, but by analyzing the structure, they
now know exactly why the antibiotics don't bind as well. They are
already using this information to try to design new antibiotics.
These kinds of protein structure studies can guide the way in designing
medicines to target defective proteins that cause all kinds of
disease.
PSI's first phase was dedicated to figuring out how to process
proteins and determine their three-dimensional structures more
efficiently. Its focus was on developing innovative approaches
and tools such as robotic instruments. Nine pilot centers created
what is now a collection of more than 1,100 protein structures,
which serve as templates for modeling related sequences. Software
developed by the PSI can now compare a structure against three-dimensional
structural templates and identify functionally important motifs.
Phase 2 is the production phase, in which thousands more protein
structures will be solved and put into the Protein Data Bank (http://www.rcsb.org/pdb/),
a public repository with powerful tools for processing protein
structure information. Some centers will also work to develop new
methods — for instance, for solving the structure of membrane proteins.
For more information, including some provocative images, see http://www.nigms.nih.gov/psi/. 
