The Protein Tango

First, the unstructured pKID is loosely bound to the folded KIX through weak, nonspecific, hydrophobic molecular interactions. Then an intermediate, partially structured complex forms as pKID writhes into shape on the surface of KIX. Eventually, pKID assumes its fully folded shape and becomes bound pKID, according to this month’s Hot Paper, in the June 21, 2007 issue of Nature.

“What Peter’s work has done very beautifully is to demonstrate, in a system, what those steps are,” says Lewis Kay, a biophysicist and leading NMR spectroscopist at the University of Toronto.

“It was the first time that transient complexes had been seen,” Wright says. “No one had any clue about how this would happen.”

Since its publication, other scientists have used the same method Wright and his colleagues employed to capture fleeting protein interactions. At Charles Craik’s lab at the University of California, San Francisco, researchers are using small molecules to target an unstructured element in the herpes virus proteases, dimeric enzymes that consist of two proteins bound together by a single helix. This unstructured element resembles the way pKID is unstructured prior to fully binding KIX. “We’re trying to disrupt this protein-protein interaction with small molecules in order to knock out the activity of these proteases,” says NMR spectroscopist Gregory Lee, Craik’s postdoc, and Wright’s findings have helped. Knowing more about how unstructured protein elements behave when contacting their binding partners “gives you clues as to which particular region to target,” Lee says, adding that they soon hope to publish a paper presenting the findings.

Protein residues are important in the progression of a variety of diseases and signaling events. In 2002, researchers found that almost 80% of the proteins in the human cancer-associated proteins (HCAP) database contained disordered regions.¹ Indeed, disordered domains, such as the tumor suppressor p53 and the kinase inhibitor p21, are likely key factors in the development of cancer and in other cell-signaling cascades, Wright says.

Detailing the mechanisms behind the behavior of these proteins—how they find, bind, and fold with their partners—could present opportunities to discover and exploit vulnerabilities in disease or signaling pathways, according to Anthony Mittermaier, an NMR spectroscopist at McGill University in Canada. “If you have an interaction between a folded protein and something that is kind of sloppy, it becomes harder to know how to start,” he says. Plus, knowing the folding process can reveal new sites for drug developers, he adds. “More information on how [disordered proteins] interact will hopefully give us a better idea of how to disrupt these interactions using small molecules.”

The tricky part of detailing the interaction between pKID and KIX was that the intermediate complexes that form through the interaction are highly unstable and evade inspection using traditional protein imaging techniques, such as crystallography. These ephemeral structures appear and vanish like ghosts. To get around this, Wright and his team used nuclear magnetic resonance (NMR) spectroscopy to characterize the highly dynamic changes in the two proteins. To interpret the data, they used an existing method—relaxation dispersion—that had previously been employed to look at enzyme activity, and isolated protein folding events. This gave them an idea of the kinetics involved in the interaction, and showed that these two protein elements were forming an intermediate state between being unbound and being bound. The methodological approach that Wright and his team used “allows us to look at details of interactions that we really couldn’t glean before,” according to Kay, who has applied the method to looking at the interaction between the protein ubiquitin and the Src-homology 3 (SH3) domain, which is critical in several signaling pathways.²

Mittermaier says that he read Wright’s paper as he was starting his own lab at McGill. “I brought [the paper] to one of our first group meetings. I held it up as an example of what we should be trying to accomplish ourselves.” He and a colleague recently used NMR combined with a calorimetry technique to show that the disordered SH3 binds a short protein “several orders of magnitude faster” than had been previously demonstrated.³

Still, it remains unclear why disordered proteins behave the way they do. Gerhard Hummer, a theoretical biophysicist at the NIH’s National Institute of Diabetes and Digestive and Kidney Diseases, says that he may have an explanation. When he ran simulations in which a fully structured pKID was made to bind with KIX, “the pre-structured peptide was binding more slowly by a factor of two,” Hummer says.⁴ This may mean that being disordered could increase the speed and efficiency with which these proteins propagate intracellular signals. “Because [pKID] does not have to fold for the binding to be initiated, the entire population is ready to bind,” Hummer says. “So there is no disadvantage to being unfolded.”

The original version of "The Protein Tango" contained a typo that suggested the protein domain pKID was able to bind to itself. The passage was meant to describe the structural changes that pKID assumes as it becomes the bound version of the domain. The Scientist regrets the error