Nature Structural Biology pp 963 - 967

Parkinson's disease (PD) is a progressive disorder of the central nervous system affecting more than 1 million people in the United States. PD is caused by the degeneration of dopamine-producing cells in the brain. Since dopamine itself cannot pass through the blood-brain barrier, the standard treatment for PD is administration of levodopa, which can reach the brain where it is converted to dopamine. A problem with this compound is that it can also be converted to dopamine in the bloodstream, leading to unwanted side effects. One way to mitigate this problem is to administer inhibitors of DOPA decarboxylase, the enzyme that converts levodopa to dopamine in the blood.

In the November issue of Nature Structural Biology, Peter Burhard of the M. E. Muller Institute for Structural Biology at the University of Basel, Switzerland and his collaborators describe the structure of DOPA decarboxylase with carbiDOPA, an inhibitor used in Parkinson's disease treatment. The structure reveals how this drug binds its target enzyme and suggests ways to design inhibitors that bind more tightly and more specifically, possibly leading to better treatments for Parkinson's and other neurological diseases.

Nature Structural Biology pp 953 - 957

In the era of high-throughput and genome-scale studies, one of the most important tasks is to categorize and correlate the huge amount of new information with what is already known. Structural genomic studies, for example, have the potential to quickly generate many new structures of proteins - so quickly that the biochemical and functional characterization of these proteins may lag behind. It is possible to hypothesize about the function of these proteins based on their structures, by examining their similarities and potential evolutionary relationships to other, well-characterized proteins. Until now, this task of assigning evolutionary relationships through structural similarity has been performed by human experts. With the increasing pace of structure determination, detailed analysis and classification by human experts will become a bottleneck in the process, and a robust automated algorithm is clearly desirable.

In this issue of Nature Structural Biology, Liisa Holm and coworker at EMBL Cambridge, UK, report an algorithm that automatically assigns evolutionary relationships through structural similarity. Importantly, the assignments from the algorithm agree reasonably well with those by human experts. This automated approach will thus be very helpful in providing functional hypotheses that can be tested by experiments. Olivier Lichtarge at Baylor College of Medicine, Houston, USA, discusses the algorithm and its implications in an associated News and Views.