Public Release: 

Duke Study Uses Tiny Temperature Changes To Probe Water-Exclusion At Protein Binding Sites

Duke University

SAN FRANCISCO -- In research that could lead to better drug design, a Duke University bioorganic chemist is exploiting the differences between two forms of water to probe how proteins' shapes control the way they bind with other molecules.

Such binding reactions, which occur constantly in watery solutions within our bodies, require removing much of the water in and around the protein's binding site to make room for the binding partner, said Eric Toone, a Duke associate professor of chemistry.

Removing this water provides much of the energy that drives a binding reaction, and determining how much energy this removal provides has previously been impossible, he added in an interview.

Toone, who made his deductions using a supersensitive instrument that can measure 1 millionths of a degree temperature changes, prepared his findings for presentation at the American Chemical Society's national meeting.

Toone's research is supported by the National Institutes of Health, the Camille Dreyfus Foundation and the Alfred P. Sloan Foundation.

Molecular binding events control many biological processes within our bodies, from the activity of enzymes to the signaling that takes place within cells. And blocking these binding events using synthetic molecules is the basis of action of almost all pharmaceutical compounds, Toone said.

Chemical reactions in the body typically unfold when smaller molecules, called "ligands" form temporary combinations, called complexes, with protein molecules at special protein binding sites. Pharmaceutical researchers can elicit biological effects by designing drugs to link up with proteins just as natural ligands do. But in the process the drugs block the natural action of the protein.

Toone said a "holy grail" of pharmaceutical chemistry -- the so-called "rational design"of drugs -- requires that scientists know enough about the architecture of proteins to devise computer programs that predict sites where the molecular linkups will occur.

"If you give the programs the protein's structure and the ligand's structure they will tell you where the ligand might bind," Toone added.

But scientists' problem is they have not yet uncovered all of nature's tricks for making the binding sites suitably "sticky," Toone said. So, surprisingly, instead of pinpointing the one actual binding site, their computer programs often show that binding is apparently possible at many different places on a protein.

Toone asked, "What's special about the actual binding site compared to all of the rest of the holes on the protein that are the right size and shape to accommodate the ligand?"

"We hypothesized that a binding site might arrange its atoms such that it's especially hydrophobic, much more so than you might have guessed by simply looking at the number and types of atoms in the binding site. That means it really, really wants to get its water out."

Under their hypothesis, "water leaving would make a much larger contribution to ligand binding than scientists had previously imagined," he added.

To prove their conjecture, Toone and his graduate students would have to analyze a complex combination of reactions involving not only the proteins and ligands but also the water that those molecules interact with.

Separating the effect of protein-ligand interactions from changes in protein-water and ligand-water interactions remains "one of the most daunting barriers to predicting how ligands bind to their protein receptors," he said.

Toone and his students were able to separate out water's involvement by using a technique pioneered in the 1950s by another professor now at Duke.

This method takes advantage of the fact that isotopes -- different forms of the same atomic element -- have very similar but not identical properties. So while hydrogen can link up with oxygen to form normal water, its heavier isotope -- deuterium -- can in turn form "heavy" water.

Normal and heavy water undergo the same chemical reactions, but heavy water interacts with other heavy water molecules about 10 percent more strongly then normal water molecules interact with each other.

This subtle difference means binding reactions in heavy water produce an extra temperature change that can be used to show how important water removal is overall to protein-ligand association.

In a 1994 report to the Journal of the American Chemical Society, Toone used this "solvent isotope effect" to show that water plays an unexpectedly crucial role in determining whether a drug will bind with a protein like a natural ligand would.

In 1996, his team followed up this finding by reporting that the way the atoms are arranged in a molecule plays a crucial role in determining how important water reorganization is to ligand binding.

If removing water from the binding site were not important, he reasoned, then the net temperature changes should be the same regardless of whether the reaction occurred in normal or heavy water. But Toone found that the temperature changes in the two solutions differed. And that meant that water plays a major part in driving the reaction.

In his new report for the ACS meeting, he will describe what he calls a novel application of the solvent isotope effect.

Using a instrument known as a "microcalorimeter," Toone and his students were able to measure temperature differences as small as a millionth of a degree as protein binding sites bound ligands in normal and heavy water.

By examining the tiny net heat changes, he concluded that by the very act of leaving the protein binding sites, water drives the binding of a ligand to a much greater extent than was previously imagined. The binding site becomes "much more hydrophobic than you would have guessed," he said.

According to Toone, the next step in his research "is to try and learn more about the relationship between shape and hydrophobicity.

"Certainly if you could find some motifs that are especially hydrophobic, to the extent that you could incorporate them into the structure of a candidate drug should help it bind much more efficiently," he said.

Much of his work in this area was done in collaboration with Terrence Oas, a Duke assistant professor of biochemistry, Toone said. He also cited valuable help from Edward Arnett, a Duke R..J. Reynolds professor emeritus of chemistry who did early work on the solvent isotope effect while at the University of Pittsburgh.


Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.