Water, water everywhere: subtly shaping protein structure and function.

For a substance essential to life, water seems remarkably commonplace: colorless, odorless, tasteless, always there. Though scientists frequently use it as a solvent and study it extensively as a molecule in its own right, they often fail to recognize that water also functions as an agent of chemical change and molecular modification in living systems.

"Water is so obvious that we almost dismiss it," admits John A. Tainer, a crystal1ographer at the Scripps Research Institute in La Jolla, Calif. "It's taken for granted."

Yet Tainer and other researchers now argue that water is no innocent bystander in the chemistry of life. True, it seems insignificant when stacked up against many of the larger, more complex molecules within a cell. But just as a swarm of gnats can drive even quite large animals to distraction, water, too, may exert a powerful effect in its environment.

In fact, water molecules don't flit aimlessly about a protein, they strongly influence its three-dimensional structure.

"Water isn't just a solvent that the proteins float around in; it has a chemical role," explains Leslie A. Kuhn at Scripps. "Water molecules are really important in catalysis and in binding." Consequently, she and others think water deserves more consideration by scientists developing drugs designed to interact with proteins.

"Water is crucial to protein folding and protein function," says David Eisenberg, a crystallographer at the University of California, Los Angeles. "But finding ways to describe the interaction between proteins and water is an elusive goal."

This dynamic interaction arises from water's polarity, A water molecule's relatively large oxygen atom draws electrons from its two hydrogen atoms, imparting positive and negative charges to the ends of the molecule. This polarity enables water to dissolve a wide array of substances in high concentrations.

Polarity also gives water the power to mold complex molecules. Often, such molecules possess both polar and nonpolar parts. The polar parts interact well with water, while the nonpolar, or hydrophobic, parts seek to avoid water. Thus, just as oil droplets mixed in water will merge into larger droplets, nonpolar parts also tend to cluster. They are, in a sense, "squeezed" together by the water around them, because water molecules prefer to associate with one another rather than with these nonpolar entities.

The polarity o! water molecules also prompts them to align with their negative ends close to the positive ends of other molecules, and vice versa. Once aligned, weak interactions called hydrogen bonds tend to maintain this formation. Water's hydrogen bonds interact with many substances, including some of the amino acids that make up proteins. In this way, water affects protein structure.

Like other researchers, Kuhn, Tainer, and their colleagues Michael A. Siani and Elizabeth D. Getzoff at Scripps want to learn more about protein structure and water's influence on it. They cannot see these molecular gnats directly, so they have developed new ways to analyze X-ray diffraction data. "They [have created] a new computational tool that will be one of a whole group of tools that modern protein chemists will use," Eisenberg predicts.

The Scripps group studied data from 56 proteins whose crystal structures had been delineated by other scientists. The proteins' various amino acids - especially their side chains (SN: 1/30/93, p.72) - give shape to a protein surface. This topography o! grooves, bulges, and bumps determines what target molecules link with each protein.

Water influences that landscape, especially when it snuggles up against the protein surface.

As proteins crystallize, their crystals trap water - in fact, water accounts for 27 to 77 percent of the crystals' volume. The Scripps researchers focused just on the 10,837 water molecules - from 56 to 690 per protein -- that had nestled close enough to proteins to show up in the X-ray diffraction data every time. These are the water molecules most relevant to protein activity, Kuhn explains.

The Scripps scientists developed special computer-implemented algorithms for analyzing the interactions between bound water and a protein surface. With one set of equations, they determine whether a molecule --water, in particular -- can touch a protein surface at any given point, Kuhn explains. The second set takes a more global look at the protein and evaluates how tightly atoms pack at a particular spot.

The first algorithm works by placing an imaginary sphere at each point on a proteins surface, then expanding that sphere until it rubs up against another part of the protein. If that point happens to exist in a groove, then the sphere can swell no bigger than the width of the groove. But at the tops of bumps, the sphere can expand quite a bit, says Kuhn. This analysis generates a map of places where water can fit on the protein surface.

The second algorithm evaluates how crowded atoms are. The computer simulates an imaginary sphere around each point on the protein surface, using that point as the center of the sphere. It then counts the number of protein atoms in that sphere and derives the atomic density around the center point. Low densities occur at bumps or protrusions; deep grooves exhibit high atomic densities, says Kuhn. The researchers can then translate these numerical data into a map of surface topography,

Basically, a protein looks like a globule, some areas of which are convoluted with deep grooves wide enough for one water molecule. The researchers found about three times as many water molecules on surfaces with deep grooves as elsewhere on the protein. Those grooves took up a quarter of the proteins surface area yet held onto half of all. the water molecules on that surface, the scientists report in the Nov. 5, 1992 JOURNAL OF MOLECULAR BIOLOGY 

Other research indicates that water molecules hop in and out of the proteins' grooves at an incredible pace: more than 1 million times in a thousandth of a second. Nevertheless, water molecules in the grooves may form chains or networks that extend throughout the protein surface. There, they can influence the fine-scale structure and function of the protein.

"The water molecules are stabilizing the surface topography," Tainer explains.

A proteins kinks and folds arise because of the particular sequence of its chain of amino-acid building blocks. Without water between them, two folds forming a groove might keep bending until they closed the groove. Thus, water keeps the grooves -- and all the amino acids inside - accessible to target molecules. "And [grooves are] where the action is:' Tainer adds.

Scientists know that side chains influence the water affinities of amino acids. Very polar side chains such as aspartic acid, glutamic acid, lysine, arginine, and to a lesser extent histidine seek a closeness with water. Leucine, isoleucine, and valine tend to avoid water molecules. But the Scripps results indicate that these affinities dominate only in grooves.

Inside grooves, water molecules seem quite picky about the amino-acid side chains they settle near. Negatively charged side chains of aspartate, for example, attracted an average of 9.3 water molecules per square nanometer, whereas valine's side chain lured fewer than one water molecule per square nanometer. Asparagine, serine, and glutamine, whose side chains both offer and accept hydrogen bonds, bind more water than amino acids with positive side chains but fewer than those with negative charges.

Outside the grooves, however, those differences fade. "They all act similarly," says Kuhn. Aspartate kept about two water molecules nearby; valine still had slightly fewer than one. "Everyone else thinks that hydrophobicity is hydrophobicity, but it actually depends on the [protein] surface shape;' Kuhn adds.

Another unexpected finding, says Tainer, is that peptide bonds, the backbone of the amino-acid chains, also interact with water molecules in the groovesas much as or even more than the most water-loving side chains.

These relationships hold no matter how big the protein. Indeed, larger proteins contain more grooves and more water -- and that water resides in the grooves, says Kuhn.

She and her colleagues have analyzed the water positions in similar enzymes found in very dissimilar organisms, including the superoxide dismutase enzymes found in yeast, cows, and humans. "Our preliminary data show that the water in grooves is conserved [across species]:' says Tainer. Such a finding adds weight to the notion that the grooves play important roles in how a protein functions.

The grooves offer drug designers attractive targets, or docking sites, for new therapeutic molecules, says the Scripps team. Often, scientists examine new drug structures by simulating the molecule with a computer. The Scripps results indicate that these simulations should include water molecules in key places, such as grooves; otherwise, the shape and chemistry will not be right, Tainer notes.

The presence of water may affect how quickly a protein can bind a drug, because the drug may first have to shoulder the water aside. Consequently, the binding part of the drug may need a stronger affinity for the proteins groove than water has. Or it may require an ability to attach to the protein with water still in the groove. Finally, the Scripps results indicate that the proteins chemistry may vary along its surface. "It's common knowledge that binding usually occurs in deep clefts or grooves:' says Kuhn. "What people haven't known is that the chemistry of water is so different in the grooves. If water is different, then you'd expect that of other molecules, too."

Thus, the Scripps findings hint that more needs to be learned about the grooves in proteins and about how water, particularly water in grooves, alters the chemical preferences of proteins and thus their binding by other substances. Such understanding, says Eisenberg, "is the essence of designing new drugs."