NEW YORK (GenomeWeb) – A team led by University ofTexas and University of Toronto researchers has generated a protein interaction map covering more than 1 million interactions across 122 species representing a broad swath of animal evolution.
Detailed in a paper published this week in Nature, the study found significant conservation of protein interactions over time, including what University of Toronto researcher and study leader Andrew Emili described as "a core network that has pretty much been preserved from the common ancestors of all multicellular life."
Such conservation, Emili told GenomeWeb, was not entirely unexpected, but, he said, to date researchers had not demonstrated it on such a large scale.
In the study, Emili and his colleagues used a combination of biochemical fractionation and mass spec analysis to identify protein interactions in five organisms: Caenorhabditis elegans (worm), Drosophila melanogaster (fruit fly), Mus musculus (mouse), Strongylocentrotus purpuratus (sea urchin), and human. The also collected data on frog, sea anemone, amoeba, and yeast that they used for independent validation of the initial interaction data.
It total, they looked at 6,387 fractions from 69 different experiments, building a network of 16,655 high-confidence interactions in humans, all of which had biochemical evidence in at least two of the species they examined and roughly half of which (8,121) were evident in three or more.
Using this experimental data, the researchers were then able to predict more than 1 million interactions expected between protein pairs orthologous to those identified via their mass spec work. They predicted interactions for 122 eukaryotes, assigning in the range of 8,000 to 15,000 interactions per species.
According to the authors, more than two-thirds of the conserved interactions they reported were novel, expanding the number of such interactions to the point where they estimated they had identified roughly 10 percent to 20 percent of the "conserved animal cell interactome."
"Certainly people have done this [on a smaller scale]," Emili said. "And the underlying hypothesis of evolution is that things are conserved — things that are in mouse will be in human, for instance. So the concept is an established one."
"But by no means has anyone looked at it systematically on such a large scale," he said, noting that the effort took some five years of fractionating and mass spec time. "That is the unique thing in this study. We actually checked biochemically how consistent interacting protein pairs have been over evolution — checked it experimentally," said Emili.
The researchers used fractionation via a variety of chromatographies to isolate proteins that consistently co-eluted, the notion being that such proteins are likely interactors. They then used mass spec analysis of these fractions to identify the co-eluting proteins.
It is possible proteins might co-elute for reasons other than that they are interactors, Emili said, noting that this is "a fundamental problem — we are not always able to discern what is truly physically associated versus what happens to just come out together."
However, use of multiple types of fractionating helps increase confidence in the results, he noted. "[Proteins] are more likely to be in a complex if they consistently co-elute together" across different chromatographies.
And, in the case of the Nature study, the researchers were given further confidence by the fact that proteins not only consistently co-eluted across different chromatographies, but across different species and spans of evolutionary time.
"The real power of our study came from the fact that we are comparing species," Emili said. "And it turns out that many proteins change their retention time on the same column across big evolutionary distances. The surface of the protein changes a little bit, so the charges, etc., tend to change over time across species.
"But the pairs typically still co-elute together," he said. "So, that is one advantage of the study — we are leveraging the beauty of evolution. [Evolution] is switching things around in terms of how [proteins] bind to the column, but the interacting proteins still consistently co-elute together, and that gets rid of a lot of the noise in the data."
He noted that while the researchers expected to find conservation of protein complexes across species going into the study, they were surprised to discover that a significant portion of the animal interactome has been "pretty much preserved from the common ancestors of all multicellular life."
"To find a network [of proteins] that is pretty much in every single animal cell — at face value there would have to be some degree of that, but the extent of it was pretty surprising," he said.
The findings have potential implications for disease research given that a substantial number of the proteins in the conserved complexes are linked to human disease, Emili said.
In particular, he noted, it could help researchers evaluate the suitability of a given model organism for understanding processes in humans.
"We usually jump back and forth in model organisms to study [human] disease," Emili said. "Now we can tell you precisely, for instance, that this complex is present in worm or in fly. So we can say that these are good model systems to explore that disease-related gene product [and its interactors]."
One thing he noted the study doesn't get at is to what extent protein interaction networks have diverged through evolution.
"That we are not able to answer because if we don't see consistency we don't know if it was because the interaction was lost or because we missed things," Emili said. "So, the flipside, how divergent things have become, that is still an open-ended question."