Understanding Protein Interactions: A Novel Approach

The Howard Hughes Medical Institute (HHMI)is reporting that investigators from Baylor College of Medicine and Harvard Medical School have devised a novel new way to study interactions of proteins in neurons from patients with ataxia. Howard Hughes Medical Institute scientist Huda Y. Zoghbi and her colleagues reported results of protein interactions (dubbed by HHMI an “interactome” project) in an article in the latest issue of the journal Cell.

To map the interaction of proteins involved in the ataxias, the researchers started with 54 proteins that were either directly involved in known diseases, or that were associated with those causative proteins. The researchers then screened those proteins for interactions with proteins produced by two huge libraries of human genes, including one library of genes specifically associated with the brain.
To detect these interactions, the researchers used the yeast two-hybrid assay–a technique invented by HHMI researcher Stanley Fields at the University of Washington. The assay involves attaching proteins to either of two separated components of a protein known as a transcription factor that switches on a telltale marker gene inside yeast cells. If the two proteins interact within the yeast cells, the transcription factor components are brought together, and the marker gene is activated.
The researchers’ screening revealed some 770 interactions between the ataxia proteins and those in the libraries. According to Zoghbi, the vast majority of these interactions had never been detected by researchers before. To test the validity of these findings, the researchers directly determined whether a sample of these interactions actually occurred in mammalian cells. Those experiments showed that more than 80 percent were valid interactions in cells. What’s more, when they explored the scientific literature for known interactions between the proteins they had identified, they found 4,796 additional protein-protein interactions.
The protein interaction network has already given the researchers insight into the function of the disease-related proteins, said Zoghbi. She and her colleagues found that many of the proteins on their map were involved in RNA transcription and processing and protein degradation. “That finding tells us these are processes that are probably very important for Purkinje cell integrity and that somehow perturbance of these processes is responsible for their degeneration,” said Zoghbi.
Another key finding, said Zoghbi, was that certain proteins seemed to be “hub” proteins in the networks – interacting with many other proteins. Studying these hub proteins could yield insight into both the basic workings of Purkinje cells, which are important to the part of the brain that integrates sensory input with motor ouput, and how the many ataxia-causing proteins cause the common pathologies that destroy them. Understanding the function of the hub proteins could yield basic insight into the machinery of Purkinje cells, she said.
Importantly, said Zoghbi, the researchers found that that many proteins that animal studies had indicated either aggravate or reduce the effects of ataxia proteins showed direct physical interactions with the disease proteins in their interaction map. Identification of these genetic modifiers is important, said Zoghbi, because they might prove effective targets for drugs to treat ataxias. “Such findings can enable us to prioritize which modifiers we should be studying,” she said. “Perhaps if we study these direct interactors, we can identify the processes that are directly impacted by the mutation and learn to affect them with drugs.”
Overall, said Zoghbi, the ataxia protein-protein interaction network will enable a far more strategic approach to studying the diseases. The technique of starting with known disease-causing proteins and developing interaction networks could yield invaluable insights into such disorders as Parkinson’s disease, diabetes, and hypertension.

The press release…
The press release at Baylor…