NEW YORK (GenomeWeb) – As technologies like next-generation sequencing continue to increase the genetic data researchers are able to generate, distinguishing between biologically relevant and irrelevant mutations and alterations has become a major issue for clinical genomics.
One potential tool for addressing this challenge is proteogenomics, the combined analysis of protein-level and gene-level information. Based on the notion that biologically significant genomic alterations will likely result in changes at the protein level, proteogenomics has become an area of keen interest within omics research, playing a key role in initiatives like the National Cancer Institute's Clinical Proteomic Tumor Analysis Consortium and Cancer Moonshot project.
In many cases, researchers have looked for alterations in protein expression stemming from genetic mutations. A study published in Nature Methods this week by scientists from the University of Graz and the Max-Planck Institute for Molecular Genetics takes a different tack, examining instead their impact on protein-protein interactions.
While changes in protein expression can be informative and are, of course, the basis of a number of clinical biomarkers, changes in protein interactions could prove even more significant, noted Ulrich Stelzl, a professor at the University of Graz and senior author on the study.
"It's clear that protein interaction is much more immediately connected to a phenotype than protein expression," he said. "Any protein in the cell is dependent on interaction. There are no proteins that are swimming around alone."
"Therefore, we went for protein-interaction mapping," he said, adding that "there is, of course, a huge gap between a nucleotide change [in DNA] and something actually happening at the protein level. The idea of this work is to close this gap a little bit with a high-resolution dissection of the effect [of a mutation] on protein interactions."
To study the effects of genetic mutations on protein interaction, Stelzl and his colleagues combined yeast two-hybrid protein interaction studies with programmed mutagenesis of the proteins being studied. They took as their model the right protein subunits of the BBSome, a human cilia protein complex involved in the genetic disorder Bardet-Biedl syndrome, which causes intellectual disabilities and various physical abnormalities.
To assess protein interactions disrupted by a given mutation, the researchers used reverse yeast two-hybrid experiments. In a conventional Y2H experiment, protein binding events bring into proximity the binding and activation domains of a transcription factor that activates transcription of a reporter gene that allows detection of a protein-protein interaction.
Conversely, in a reverse Y2H experiment, the reporter gene is turned off in the event of a binding event, meaning that positive interactions are indicated by a lack of signal.
To assess changes in interactions due to mutations, Stelzl and his colleagues first ran a conventional Y2H experiment to determine the interactions present. They then introduced the mutation of interest and ran the reverse Y2H experiments, looking for interactions that were disrupted by the mutation.
Key to this process, Stelzl noted, was the development of a new synthetic promoter that reduced the problem of background yeast growth that has traditionally proven a challenge for reverse Y2H experiments.
Reverse Y2H techniques have existed for around two decades, but they have not been scalable to large screening formats, Stelzl said.
"We engineered this promoter construct to optimize the procedure," he said, noting that the challenge was achieving a construct that was repressive enough to limit background yeast signal while not so repressive that it suppressed all yeast growth.
Using the approach to mutate every amino acid in the eight BBSome proteins and screen for changes in their interactions, the researchers identified more than 1,000 amino acid mutations that led to a disruption of an interaction.
They then took these mutations and highlighted those known to occur in cases of different diseases.
"These are the interesting ones," Stelzl said. "Because it essentially means that these [are the disease-linked] mutations that would change a protein complex, and so there is a good chance that it is actually causal."
Y2H looks at individual protein-protein interactions, as opposed to full protein complexes, which raises the possibility that the approach might not account for compensation by other complex components for a given mutation or the effects of multiple mutations within a single complex, Stelzl said. Nonetheless, it provides a starting point for identifying alterations likely to be worth following up on.
In theory, researchers could apply the approach on a proteome-wide scale, he added. "Then you would have these [mutational] profiles for each and every protein, and if there is anyone who has a mutation and you want to know the potential effect, you could go to these profiles and see whether it was an impactful mutation or not."
Based on Y2H and NGS technologies, both of which are commonly used for genome- and proteome-wide analyses, the approach should be amenable to large-scale analyses, Stelzl said. The limiting step is most likely to be creating the mutant gene pools for such studies, he suggested, adding that this was more a matter of cost than technology.
Stelzl said that he and his colleagues have since expanded the work beyond the BBSome and are looking at various systems linked to certain oncogenes and tumor suppressors.
"I think it could be highly valuable for assessing certain cancer mutations and their effects," he said. "If you think about some tumors, they have, say, 10,000 mutations, including not very common ones. And if you have to seek out which ones are potentially of interest, then you can say if they cause a change in an important interaction that is maybe in an important pathway, they you could prioritize those and see if they could be targeted [therapeutically]."