Leveraging Disorder

ASK BIOCHEMISTS to catalog their major benchtop aggravations, and unstructured proteins are sure to make the list. These proteins have long sections of amino acids that flail about in ways that cause protein aggregation, messy nuclear magnetic resonance (NMR) spectra, and general malaise in those inquisitive enough to study them. But there's more to these unruly proteins than the vexation they cause.

Many of these floppy, unstructured proteins are lead actors in key cellular processes ranging from cell division to gene transcription. In fact, by some tallies, at least 30% of human proteins are intrinsically disordered. "A lot of intrinsically unstructured proteins are also involved in disease processes, from cancer to protein-folding diseases like Alzheimer's and Parkinson's," says Richard Kriwacki, a professor of structural biology at St. Jude's Children's Hospital in Memphis.

Intrinsically disordered proteins may have a bit of localized structure-an α-helix here, a β-strand there-and they may even boast a few folded domains. Their hallmark is a region of at least 50 amino acids that just can't seem to settle down—until they find a protein binding partner. A growing number of researchers believe that this intrinsic disorder is essential to function: It confers plasticity and flexibility to proteins, thereby allowing the signaling networks these proteins control to rapidly respond to the cell's environment.

At first glance, intrinsic disorder goes against the biochemistry dogma that a well-defined three-dimensional structure is absolutely essential for protein's biological activity. Proponents of disordered proteins don't negate the biological importance of structure, however; they simply point out that it is not always necessary for function. "Proteins use all of conformational space to function," says Peter Wright, a biochemist at Scripps Research Institute. Wright and others postulate a continuum of fully operational proteins whose backbone conformations range from well-structured to fully disordered.

"We've all seen a lot of pretty protein X-ray structures. It's easy to think, 'That's how all proteins work,'" says David Eliezer, a biochemist at Cornell University's Weill Medical College. "People have mistakenly thought that the interesting parts of proteins are the same as the structured parts. This is just not true."

Protein disorder, like protein structure, is encoded by the protein's amino acid sequence, says A. Keith Dunker, a theoretical biologist at Indiana University, Indianapolis. Unstructured proteins tend to have lots of hydrophilic amino acids—including arginine, lysine, glutamine, proline, and serine—and few hydrophobic residues. In fact, the paucity of hydrophobic groups is a major reason these proteins can't fold: There's no meat for the hydrophobic core needed to form a structured protein. The high incidence of charged amino acids further destabilizes any latent tendency toward structure.

ALTHOUGH INTEREST in protein disorder has been slowly blossoming over the past decade and downright exploded in the past couple of years, the intrinsic disorder of some proteins such as casein and serum albumin was known as early as the 1950s. "We knew about unstructured proteins, but we just didn't know what to make of them," Wright says. "I didn't think much about unstructured proteins at all until a postdoc showed me a lousy NMR spectrum of what turned out to be a fully functional protein."

That was in 1996, and the postdoc was Kriwacki. Last month, Kriwacki coorganized the inaugural meeting of the newly minted Intrinsically Disordered Protein Subgroup at the Biophysical Society's annual meeting in Baltimore. "Today, we mark the beginning of this field being solidly recognized," Kriwacki told a standing-room-only crowd of about 150 people.

Biologists mostly ignored the value inherent in disorder until a few milestones happened, says Kriwacki. First, the sequencing of multiple genomes spawned databases of genome sequences. Later, bioinformaticians developed algorithms to predict from such genome data which proteins contain segments of intrinsic disorder. Using these techniques, researchers have built a comprehensive inventory of intrinsically disordered proteins, which lists a who's who of biologically significant proteins. When Wright and molecular biologist Jane Dyson of Scripps published a review of important intrinsically disordered proteins in Nature Reviews of Molecular Cell Biology in March 2005, "the broader scientific community started to get on the bandwagon," Kriwacki says. In fact, since 2002, the number of papers published annually in this field has more than tripled to over 450 a year, Dunker says.

As more researchers tune into intrinsic disorder, the perks of being amorphous are coming to light. One major benefit is that disorder permits protein promiscuity: In other words, disordered regions of proteins are not limited by their structure to a single relationship with a single binding partner.

"Proteins are social creatures. They have a variety of different relationships with different proteins, much like humans have with other humans," says Péter Tompa, a biochemist at the Hungarian Academy of Sciences in Budapest. "A person can be a father to some, a friend to others, and a colleague to yet another. For proteins, it's the same."

By sidestepping structure, disordered regions have the flexibility to bind an array of different protein partners, allowing the same sequence to multitask. Furthermore, such structural suppleness explains how a protein such as p53, the infamous regulatory protein that is often implicated in cancer, can bind up to 1,900 different proteins in the cell.

Indeed, a majority of p53's protein-binding sites are located within the 29% of its amino acid sequence that is intrinsically disordered, Dunker says. For example, a short 10-amino acid sequence on p53's unstructured C-terminal end binds at least four entirely different protein partners, affecting processes ranging from gene expression to cell division.

The act of binding, however, forces these promiscuous regions to customize their structures depending on their binding partner. The same short sequence can bind in an α-helical conformation, a β-strand conformation, and two other conformations, all depending on the binding partner.

Intrinsically disordered proteins also are more likely than more structured proteins to be "hub" proteins in the behemoth complexes that regulate cellular processes such as transcription. Long, unstructured linkers between globular domains act as a flexible scaffold for these complexes.

The extended nature of an unstructured protein is useful for more than simply recruiting many players into a large protein "party." Disorder permits these proteins to wrap around their structured binding partners, making up to three times more surface area contact than is available between two well-structured proteins.

THE PERK of a dramatically larger surface area of interaction often comes at the cost of entropy, however. Every time an unstructured protein binds its partner, the universe becomes just a bit more locally ordered. The cost-benefit balance means that some unstructured proteins have relatively low affinity for their protein partners compared with structured proteins, although there are some exceptions.

This low affinity provides a plasticity that biology certainly exploits. When cells have to react quickly, strong transient interactions make it hard for protein-protein complexes to quickly fall apart, a necessary trick in a cell's rapid-response toolbox. Instead, biology often uses a multitude of transient, low-affinity interactions among many protein partners to provide the specificity that a cell needs for tightly orchestrated processes such as gene transcription or cell division.

Many of these tightly regulated processes are turned on or off by phosphorylation or acetylation. These chemical modifications of specific residues act as switches to target some proteins for degradation or to activate entire protein signaling pathways. Many amino acid residues that get modified in this way are located in highly unstructured regions of a protein, Kriwacki says.

The phosphorylation of an intrinsically disordered region in p27 that activates mitosis exemplifies how lack of structure promotes function.

Kriwacki's group recently showed that when p27 wraps around its protein partners Cdk2 and cyclin A like a two-pronged staple, mitosis doesn't occur until phosphorylation in an unstructured region of p27 kicks out one of its prongs. "The flipping out of the prong from the active site depends entirely on p27 being a flexible protein," Kriwacki tells C&EN.

Once p27's prong is flipped out, a threonine residue becomes open to phosphorylation. If it is phosphorylated, p27 is marked for degradation, which then allows mitosis to occur.

"Phosphorylating one site can cause a chain reaction that acts to influence the phosphorylation of a second, and possibly a third or fourth, phosphorylation site," Eliezer comments. "This is turning out to be a common mechanism," and not just in cell division, he notes. For example, Eliezer's group is currently looking at how the protein α-synuclein, which has been implicated in Parkinson's disease, alternates between a highly unstructured and a structured state subject to phosphorylation of multiple sites in disordered parts of the protein.

Indeed, neurodegenerative diseases such as bovine spongiform encephalopathy (mad cow disease) and Parkinson's, where unstructured protein segments take part in an aggregation that leads to the formation of potentially harmful fibrils, highlight the biological cost of intrinsic disorder. But researchers such as Dunker have shown that many proteins associated with diabetes, cancer, and cardiovascular disease also have regions of intrinsic disorder that drug companies may want to seriously consider. In fact, Dunker points out, because many pharmaceutical companies have major research programs directed toward protein phosphorylation, there has already been "an indirect focus on intrinsically disordered proteins in drug discovery."

"Biochemists should realize that disordered segments of proteins often play the most critical roles in function," Kriwacki says. Teachers should add a little disorder—of the protein kind—to their curricula, adds Gary Pielak, a chemist at the University of North Carolina, Chapel Hill, because "disordered proteins are so different from what students learn in class."

The key, Kriwacki says, "is to always consider 'unstructure' in relation to function."