Scientists at the European Molecular Biological Laboratory and the Technical University of Denmark have combined microarray data with protein-protein interaction data to come up with a novel way of observing how protein complexes form dynamically, with a few elements of the complexes corresponding temporally to the cell cycle.
“Past studies of this type have usually left out a crucial element — time,” said Peer Bork of EMBL who led the research group, whose study is published in this week’s issue of Science. “But now a picture has emerged which is extremely dynamic.”
To get a dynamic picture of protein complex formation in yeast, Bork and his research group integrated several different data sets that had been produced in the late 1990’s. One data set measured the expression levels of proteins at certain times during the yeast cell cycle. Other data described protein interaction networks, which have been studied for a long time in Bork’s laboratory.
The researchers discovered that in yeast, key components needed to create a functional complex are produced ahead of time and kept in stock within the organism. When the complex is needed, a few crucial last pieces are synthesized, and then the apparatus is assembled. Holding off on the last components allows the cell to keep from building the complexes at the wrong times. This is a different scenario than in bacteria, which typically start producing all parts from scratch when they want to form an active complex.
“Lots of bioinformatics were needed to measure for each gene in yeast if it was periodically expressed or not, and if it cares about the cell cycle,” Bork told ProteoMonitor. “Then if you map the cell cycle proteins to networks, you get a much better picture of the functional cascades.”
Bork said that his research on protein complex formation does not have any direct consequence on medical or drug research, but that it indirectly impacts on drug discovery by revealing a deeper picture of how active complexes form.
“You could have 10 subunits of the complex, and one subunit that corresponds to the cell cycle is enough to make the thing happen,” said Bork. “When a particular subunit is added and expressed could impact on the thinking about how drugs should target complexes.”
Peter Lu, a senior research scientist at the Pacific Northwest National Laboratory who uses single-molecule photon-stamping spectroscopy to study dynamic interactions of subcellular components, said that Bork’s research is “pretty novel in terms of putting time into the study.” Understanding the dynamics of protein-protein interactions is definitely important for drug design, he added.
A drug can be effective in two ways, Lu explained. “One way is that it may be thermodynamically strong so that it may not immediately bind, but when it does, it’s thermodynamically tight. Another type of interaction may be kinetic. The complex may not be very stable, but the kinetic activation barrier is low.”
By understanding the kinetics of protein complex formation, scientists can better design drugs that would be kinetically favorable to binding, said Lu.
Lu noted that one element that was not addressed in Bork’s study was the quantitative characterization of protein complex components.
“This research is at an early stage,” said Lu. “I think in the future, they will follow up with more specific quantitative characterization and a more sophisticated study will appear.”
Gordon Hager, a researcher at the National Cancer Institute who studies dynamic changes in protein-DNA interactions using ultraviolet laser, pointed out another element that was missing in Bork’s paper — the reference to recent work on real-time dynamic movement of proteins.
“What we need is common terminology between the old world of footprinting and the world of chips and now this world of real time experiments,” said Hager. “They should have at least referred to the growing literature from people who look at protein movement in real time.”
Bork said that in the future, his group would like to compare protein interaction data with other cell cycle data, including data from humans. The researchers would like to push beyond complexes to entire interaction networks, he added.
In addition, the researchers would like to study the dynamics of how protein complexes degrade.
“It’s equally important to see when a complex degrades,” Bork pointed out. “At this point, we have data on degradation of proteins, yes, but we’re not aware of any data which would do the job [of providing a dynamic degradation picture].”
Aside from providing a dynamic picture of protein complex formation, the current yeast study also identified some new molecules in the complexes that had been overlooked in the past, said Bork.
“We found some new periodically expressed proteins that have so-called degradation signals — that is, a phophoresidue that signals for the thing to get degraded,” said Bork.
Studying how post-translational modifications, such as phosphorylations, are temporally regulated might provide an indirect model of degradation, Bork said.
Bork speculated that it is costly in terms of energy for an organism to control things temporally. Therefore, it is more efficient for most components of a complex to be expressed all the time, or without a pattern of periodical expression, and for only a few key elements to be temporally regulated.
PNNL’s Lu said that it is not clear if time-coordinated processes take more energy.
“That is quite a deep question to ask at the moment,” he said. “It’s an interesting question to ask, and I think that still needs to be studied.”
Aside from Bork, other researchers who contributed to the study of dynamic complex formations in yeast were Ulrik de Lichtenberg and Soren Brunak of the Technical University of Denmark, and Lars Jensen of EMBL.